So, you want to become a Data Scientist.

Maybe you’re pivoting from

a different eld, or maybe you’ve just graduated and you’re looking
at this mountain of skills to learn and con icting information about
how to prioritize your time, and you’re wondering, “Where do I even
begin?”
Don’t worry. In this roadmap, I’m going to walk you through the
exact steps to go from zero experience to landing your rst role as a
Data Scientist in just one year – without a bootcamp or advanced
degree.
You’ll learn what skills to prioritize (and in what order), how to build
real-world experience before you even have a job, how to make your
resume and LinkedIn stand out, and how to excel in tricky technical
interviews. We’ll even talk about what life is like in your rst role so
you know what to expect and how to get your career started on the
right foot.
Quick intro: I’m Marina. I work as an Applied Scientist (a hybrid between a Data
Scientist and a Machine Learning Engineer) at Twitch (i.e. Amazon). I also mentor at
an machine learning bootcamp and coach people one-on-one on how to break into
the eld.
I transitioned into data science from a non-technical background, so I know rsthand
how challenging—and rewarding—this path can be. As a little motivation, I earn ve
times what I did before transitioning into the eld, while at the same time having
better work-life balance and a more engaging day-to-day now than I did back then.
With consistency and a solid plan, you can do it, too!

If you feel like you need more support, you can book a 1:1 call with me on topmate.io.
And don't forget to check out my YouTube channel for more Data Science education
and motivation.
Also, no pressure at all, but if you nd this guide helpful and feel so inclined, you can
click here to buy me a coffee.
Should You
Still Become
a Data
Scientist?
Right away, let’s address a question you
might be worried about: Is it even worth it
to become a Data Scientist now? With all
the layoffs in tech and the increasing
presence of AI, it’s natural to question
whether data science is still a smart career
choice.
Here’s why I believe that it is: First, while
tech layoffs have absolutely been real, they
don’t tell the whole story. Even with recent
cutbacks, data science remains one of
the most in-demand elds. According to
the U.S. Bureau of Labor Statistics,
employment of Data Scientists is
projected to grow 36% from 2023 to 2033,
which is much faster than most other
professions.
And what about AI? It’s true that AI is
reshaping industries, but it’s not
eliminating Data Scientist roles – at least
not yet. More importantly, the core skills
you’ll develop—like programming,
statistics, and critical thinking—are
valuable across many industries, even if
you do need to pivot in the future.
So, while it might be harder to break into
data science now compared to a few years
ago, the rewards are absolutely still there
for those who are persistent and focused.
If you’re willing to put in the effort, this
path can lead to a well-paying,
intellectually rewarding career with plenty
of opportunities for growth.
In This Roadmap
Timeline
Area of Focus
What to Learn
Things NOT to Learn
Marketing Your Experience
Interview Prep
How to Do Well When You Start Your First Job
Timeline
You’ve probably seen videos or courses promising you can become a Data Scientist in
just a few weeks or a few months.
While that might be realistic for someone with a super strong technical background
and a lot of free time, for most beginners, it’s not going to be feasible.
Data science isn’t easy. It’s not a job where you just learn a few simple things and
then coast. Building the necessary skills takes time, but it’s worth the investment.
This process is de nitely a marathon, not a sprint.
That being said, if you can dedicate about 20 hours a week to learning, this 12-
month timeline will guide you step-by-step from beginner to that rst job.
So if you’re starting in January, this timeline ts perfectly. If not, don’t worry, just use it
as a exible template. Let’s break it down.
January to
June
(Weeks 1-
26)
Building the Basics
The rst six months are all about establishing a strong foundation.
During this phase, you’ll learn the essential skills every Data Scientist
needs, while also starting to build your network.
Step 1 Join the Data Science
Community
First things rst, we’re going to start meeting people. Find local
meetups, Discord servers, or LinkedIn communities where Data
Scientists connect.
Building relationships early will help you down the road, whether it’s
nding mentors, collaborators, or even job opportunities. Plus, you’ll
start learning the language of the eld and getting familiar with the
kinds of things people are working on, which will be helpful when
you’re coming up with projects and preparing for interviews.

Step 2 Start Learning

This is your “reading and practicing” phase. We’re going to start with
beginner-friendly courses, books, and YouTube tutorials to dive into
the fundamentals. Throughout this phase, we’ll work on follow-
along projects in courses, as well as small self-directed projects to
apply what you’re learning.

Step 3 Showcase Your New Skills

Around halfway through this rst phase (say, around March or April),
you should be con dent enough with the basics to showcase your
new skills and start getting a feel for the job market.
Update your LinkedIn and resume with the skills and projects you’ve
worked on so far, and start applying to jobs casually—not necessarily
with the goal of landing a job right away, but to practice and get
feedback. Every application and (potentially interview) will teach you
something.
July (Weeks 27-31)
Your First Big Project
After the rst learning phase, it’s around July. Now that you’ve got a
solid foundation, it’s time to tackle your rst self-directed end-to-end
project. Think of a project that demonstrates your skills
comprehensively, such as:

• Scraping news websites and running sentiment analysis about a

particular topic to track trends over time
• Predicting a genre of a song based on its lyrics
• Building a recommendation system for something that interests
you, like board games or podcasts
• Creating a real-time Crypto price prediction model

Your goal at this point is to stretch yourself to apply your skills to

new problems that you can add to your resume and speak
con dently about in an interview.
August (Weeks
32-35)
Real-World Experience
Now that you have some experience working on
projects independently, it’s time to try to get some
experience working with real clients. Start reaching
out to family friends, local small businesses,
nonpro ts – anyone who might work with you –
and offer your data skills for free.
For example, you might volunteer to analyze
customer data for a local shop, or build a
dashboard for a nonpro t. These real-world
projects will boost your portfolio and build
connections. We’ll talk a lot more about how to do
this in the How to Get Experience section.
We’ll also want to ramp up job applications at this
time. But, we’re not going to just apply through job
boards. Instead, we’re going to proactively reach
out to recruiters and hiring managers – more on
this in the How to Get Interviews section.
Lastly, we’ll use this time to start building a habit of
practicing coding problems (LeetCode and SQL) so
that we’re ready when it’s interview time.
September
& October
(Weeks 36-
44)
The Grind
By September we should be working on a second or third major
project. Ideally this should be for a real client (even for free), but if
you don’t hear back, continue to work on larger scale self-directed
projects that simulate real-world use-cases as much as possible.
November (Weeks 45-
48)
Interview Prep Mode
By around November, your skills and portfolio should be solid, so it’s
time to shift gears to interview prep. This is something we can do
with or without a speci c interview lined up.
In addition to the coding practice we’ve been doing, we’ll also want
to start thinking about preparing for case studies and presenting
our work in the best possible light. I’ll talk about interview prep in
more detail later on in the roadmap.
December (Weeks 49-
End of year)
The Finish Line
December is when all your hard work comes together. Stay
consistent and keep pushing until you land that rst role.
At this stage, you should still be applying to jobs daily, leveraging
your network, and keeping your LinkedIn active with posts about
your projects.
You’re also going to want to continue working on your portfolio,
practicing coding challenges, and preparing for interviews.
With persistence and the right strategy, you’ll land your rst data
science role—and start the next year ready to hit the ground
running.
Area of Focus
One of the rst decisions you’ll need to make on your journey is
understanding the type of data science role you’re aiming for. As a
beginner, you might not know that Data Scientists can do very
different things at different companies.
In my experience, there are three main areas of focus: Analytics,
Data Science Generalist, and Machine Learning/production-focused
roles.
Each has its own unique set of skills, day-to-day tasks, and career
paths. Let’s break down what each one involves, and some pros and
cons.
1. Analytics
If you enjoy diving into data to help businesses make better
decisions, an Analytics role might be the right t. These roles focus
on extracting and interpreting insights rather than building
complex machine learning models. You’ll often work closely with
stakeholders to solve business problems through dashboards,
reports, and A/B tests. These roles are generally more collaborative
and less likely to have signi cant heads-down time.
Typical Tasks
• Analyzing trends in sales, customer behavior, or operational
ef ciency.
• Creating dashboards and visualizations to present data clearly.
• Running experiments to optimize business strategies.

Pros
• These roles often have a strong business impact and can be
rewarding if you like problem-solving in that context.
• Typically have a shorter learning curve for beginners since they
rely more on data wrangling and basic statistical skills.
• Great foundation if you’re eventually interested in a broader data
science role.

Cons
• Can be less technically challenging. If you’re looking to dive into
machine learning or advanced statistical modeling, this may not
be totally satisfying for you.
• Compensation can be slightly lower than in more specialized
data science or ML-focused roles.
2. Data Science Generalist
Like the name implies, data science generalist roles require a mix of
skills, from data wrangling and statistics to building predictive
models. If you enjoy variety and exibility, this path allows you to
work across the entire data science pipeline—from exploratory data
analysis to deploying machine learning models.
Typical Tasks
• Cleaning and analyzing data to uncover patterns.
• Building and testing predictive models.
• Communicating ndings through visualizations and reports.
Pros
• Offers a wide range of learning opportunities since you’ll touch
every aspect of the data science work ow.
• Opens up multiple career paths, allowing you to specialize later
on in analytics, statistical modeling, or machine learning.
• Often requires less technical depth in math and ML but still gives
you a solid, broad foundation.
Cons
• You’ll need to be exible and comfortable working on different
types of projects.
• The breadth of skills required can feel overwhelming at times,
especially if you’re just starting.
3. Machine Learning-Focused
If the technical and engineering side of data science sounds fun to
you, an ML-focused role might be the best t. These roles often
require deeper programming, math, and algorithmic knowledge,
making them more challenging in some ways, but also (in my
opinion) really fun.
These roles are more likely to allow you to have solo deep work time
and are generally less collaborative than generalist or analytics roles.
Typical Tasks
• Developing machine learning models.
• Deploying models to production systems.
• Potentially working on advanced algorithms for computer vision,
natural language processing, or applications like recommender
systems.
Pros
• Highly specialized skill set that’s in demand, especially in tech-
driven companies, and especially right now.
• Often offers higher compensation due to the technical
complexity and lack of talent to ll open roles.
• Opportunity to work at the cutting edge of technology.
Cons
• Steeper learning curve, especially for those without a technical
background.
• Requires more engineering-type work, which may or may not
appeal to you.
Which One Should
You Choose?
For this roadmap, we’ll assume you’re aiming for a Data Science
Generalist role. It’s a great starting point because it provides
exibility and opens doors to both analytics and machine learning
positions later.
That said, the advice in this roadmap applies across all three paths.
The key difference lies in what you prioritize in your learning:

• For Analytics roles: Focus more on SQL, data visualization tools

(like Tableau or Power BI), and business problem-solving.
• For ML-focused roles: Dedicate extra time to programming,
advanced math, and machine learning algorithms. For more
senior roles, you’ll also need to understand ML System Design.
The important thing is to start somewhere. As you gain experience,
you’ll naturally discover which path excites you the most.
With your focus decided, let’s move on to the core skills you’ll need
to succeed.
 What to Learn
There are different paths to learning the skills to become a Data
Scientist. You could go the traditional route with a degree, join a
bootcamp, or teach yourself.
For this roadmap, I’m assuming you’re going the self-taught
route—it’s often the fastest, the most exible, and obviously the
least expensive. But even if you’re pursuing a degree or
bootcamp, this guide will help you ll in the gaps. Because trust
me, no matter what formal program you’re in, there’s always
extra self-study involved.
That said, having an advanced degree can help. It won’t
guarantee you a job, but it won’t hurt either—resume scanners
love seeing Master’s and PhDs, and without them you might
struggle to be considered for some roles. Nowadays there are
many low-cost part time online programs, so if traditional
education appeals to you, it's something to consider.
Still, some Data Scientists break into the eld without an
advanced degree, so don’t let this hold you back.
Should You
Get
Certi cates?
There are many paid online certi cate programs available for Data
Science. While many of them are high quality and the certi cate
certainly can't hurt your chances of getting hired, don't feel
stressed if you can't afford them. You can absolutely demonstrate
your skills through self-directed, project-based learning as well, at
zero cost.
Now, a big question is where to start: With math or coding.
Here’s my take—start with both and use each to learn the other.
I think this is a good approach because they’re equally essential, and
if you don’t enjoy either, this might not be the eld for you, and it’s
important to gure that out before you dedicate a ton of time to
something you don’t actually like doing.
But don’t worry, you don’t need to master calculus or become as
good at coding as a Software Engineer right away. The idea is to
start with the basics of both and build them up together as you go.
Here’s how we’re going to do it:
• Spend your rst month learning the basics of both math and
coding. For example, study Python for data analysis while
brushing up on essential statistics.
• As you progress, pair them together. When you learn about a
concept like regression in statistics, implement it in Python. This
integrated approach will make the material more practical and
easier to understand.
A Note on Effective
Studying
Data science concepts can be challenging to understand. AI tools
like ChatGPT can be super helpful for learning, debugging, and
clarifying concepts—but don’t let them do all the work.
The straining, slightly uncomfortable feeling of wrapping your
brain around a concept is exactly how you grow. If something’s
hard, lean into it. Use ChatGPT to help, but always aim to understand
the why behind the answers, and let yourself struggle a bit before
you just ask ChatGPT to explain it to you right away. Because
sometimes it won’t be able to, and you need to have the skill and
mental endurance to gure things out on your own.
With that mindset, let’s jump into the speci c skills you need to
prioritize, starting with Python.
Python
Python is the main coding language used for data science, and for
good reason—it’s versatile, beginner-friendly, and it has a massive
ecosystem of libraries tailored for data analysis and machine
learning.
As a Data Scientist, you’ll spend a lot of time coding, whether it’s
cleaning datasets, building models, or automating work ows. While
some companies may use R, Python dominates the eld. The good
news is that once you’ve mastered Python, transitioning to R if you
need to—or any other language—isn’t that hard.
To get started with Python, we’re rst going to learn the very basics:

• Variables, data types, loops, conditionals, and functions.

• Understand how to work with data structures like lists,
dictionaries, sets, and tuples.
• Know how to handle les (read/write operations) and exceptions.
Next, we’ll need to understand package management systems like
pip and conda to install and manage dependencies for your
projects.
Then, we’ll want to get familiar with the main libraries used by Data
Scientists:
• Pandas: For data manipulation and basic plots.
• Numpy: For numerical computations.
• Scikit-learn: For building machine learning models.
The goal here isn’t to memorize every function but to become
comfortable with the tools and learn how to reference
documentation effectively when you’re stuck.
Learning Resources
Don’t get too hung up deciding which Python course to take. Any
reputable beginner course will do—just pick one and stick to it.
Some great options include:
• Code Academy
• Python for Data Analysis e-book
• Udemy courses (wait for sales/use coupon codes for Udemy)
Bonus
The above is enough to get started. But if you want to stand out, I
highly recommend the book Software Engineering for Data
Scientists (af liate link) to level-up your coding and software
development skills.

Timeline
Plan to spend about four weeks mastering the basics of Python that
we went over.
Example Projects
To solidify your learning, here are some example projects to put
things into practice:
Write a program to process words in a text le
• Read in some data.
• Count the frequency of each word and identify the top 10 most
frequent words.
• Plot the top words.
• Save the results into a new le.
Create a basic calculator tool
• Write functions for addition, subtraction, multiplication, and
division.
• Take user input and perform the selected operation.
• Add exception handling to manage invalid inputs or division by
zero.
• Print the results.
Probability & Statistics
At the same time, we’re going to start studying statistics. As a Data
Scientist, you’ll use statistics in nearly every aspect of your work—
from analyzing datasets to evaluating machine learning models, and
even designing experiments. The better your grasp of statistics, the
better your ability to interpret results and not make huge mistakes
that cost the company millions!
Just kidding. Kind of. We really do want to know what we’re talking
about when we make decisions based on data.
Just like with Python, start with the basics and build your
knowledge over time. Here’s a roadmap of some key topics:
1. Descriptive Statistics
• Understand measures of central tendency: mean, median, and
mode.
• Learn measures of dispersion: like variance and standard
deviation.
• Explore common distributions like the normal distribution,
skewed distributions, and uniform distribution.
2. Data Characteristics
• Differentiate between continuous vs. discrete variables and
nominal vs. ordinal data.
• Learn how to identify and handle outliers.
3. Correlation and Multicollinearity
• Understand the difference between correlation and covariance.
• Learn about multicollinearity and its implications for regression
models.
4. Inferential Statistics
• Understand hypothesis testing, including p-values, z-tests, t-tests,
and chi-squared and ANOVA.
• Dive into basic probability concepts, including Bayes’ Theorem
and randomization.
• Understand con dence intervals.
• Learn about sampling techniques, resampling methods, and
common statistical biases.

5. Experimentation Basics
• How to determine sample size.
• How to decide how long to run the test.
• Learn about the importance of randomization in experiments.
Bonus
Explore causal inference techniques to evaluate cause-and-effect
relationships when we can’t run an experiment.
Learning Resources
• This Udemy course is a super gentle introduction (wait for
sales/use coupon codes for Udemy)
• StatQuest (YouTube channel): Fantastic for breaking down
complex topics in an easy-to-understand way.
• Khan Academy: Offers a structured and beginner-friendly
approach to core statistics topics.
• Practical Statistics for Data Scientists book (af liate link): A great
hands-on resource that ties statistical concepts directly to data
science applications in R and Python.

Timeline
Dedicate three-four weeks to studying statistics alongside Python
(so, it’ll take around eight weeks total to learn the basics of both).
Spend the rst week focusing on descriptive statistics and data
characteristics, then move on to inferential statistics and
experimentation in the following weeks.

Example Projects
• Download a Kaggle sales dataset and analyze the distribution of
sales in each region and over time.
• Download a survey dataset from Kaggle and analyze stats for
question answers, nd outliers, and compare answers across
groups (demographic or age etc.)
SQL
Next, we’re introducing another coding language: SQL. No matter where you work,
you’ll almost certainly use SQL to interact with data, because this is the language we
use to query data from a database. While some smaller organizations may rely only
on CSV les, most companies use relational databases to store their data, and SQL is
the primary tool for accessing and manipulating it.
Luckily, it’s pretty easy! Especially if we just focus on the core concepts we need to get
started in data science:
1. Core Querying Skills

• Learn how to write queries using SELECT, WHERE, ORDER BY, and LIMIT.
• Understand how to deal with NULL values effectively to avoid errors or unexpected
results.
• Get comfortable with handling dates and timestamps.

2. Intermediate Features

• Master JOINS: Understand the differences between INNER JOIN, LEFT JOIN, RIGHT
JOIN, and FULL OUTER JOIN.
• Learn to use UNIONs and self-joins for more complex queries.
• Explore CASE statements to add conditional logic to your queries.
• Understand GROUP BY and aggregate functions (e.g., COUNT, AVG, SUM, MIN,
MAX) for summarizing data.

3. Advanced Techniques

• Dive into Common Table Expressions (CTEs) to break down complex queries into
manageable pieces.
• Learn how to write and optimize subqueries.
• Explore window functions like RANK, ROW_NUMBER, and LAG to perform
advanced analytics within a dataset. This will almost de nitely come up in
interviews, so don’t skip it.

4. Query Optimization

• Learn basic strategies to make your queries run more ef ciently.

Learning Resources
• Any intro YouTube course or Udemy Course. I took this one on
Udemy.
• CodeAcademy
• W3Schools
• SQLBolt
• LeetCode

Timeline
Dedicate two weeks to learning the basics of SQL.

Example Projects
For SQL, just work through practice problems in the interactive
hands-on learning platforms linked above.
Tableau/Power
BI/Data Visualization
As a Data Scientist, a lot of your job is about telling a story with data.
Data visualization tools like Tableau, Power BI, and Python data
visualization libraries allow you to transform raw data into insights
that non-technical stakeholders can easily understand.
While you don’t need deep expertise in these tools to get hired,
being familiar with them and having them listed on your resume is
useful.
For your rst role, focus on the fundamentals of data visualization,
both with dedicated tools and programming libraries:
1. Basics of Tableau and Power BI
• Connect to a dataset and explore the interface.
• Create basic charts like bar plots, line graphs, and scatter plots.
• Combine visuals into dashboards and add interactivity (e.g., lters
or slicers).
2. Python Visualization Libraries
• Generate static and interactive visualizations for data exploration.
• Customize plots with titles, labels, and legends.
Learning Resources
• Intro to Tableau
• Intro to Power BI
• Intro to Python visualization libraries

Timeline
Spend one week learning the basics of Tableau, Power BI, and Python-based
visualization libraries.

Example Projects
1. Sales Dashboard

• Use a sales dataset from Kaggle.

• Create bar charts to show sales performance by category.
• Add line charts to track sales trends over time.
• Use lters for interactive analysis by region or time period.
• Highlight top-performing regions and products.

2. Customer Segmentation Visualization

• Use a customer dataset with demographics, purchase history, or behavioral

metrics.
• Visualize clusters of customers using scatterplots (e.g., spending vs. frequency).
• Create demographic breakdowns using pie charts and bar charts.
• Add slicers to explore different customer segments.
• Add a KPI card showing key metrics like average spend per customer.

3. Movie Reviews

• Use a movie dataset.

• Create histograms to show the distribution of ratings.
• Use scatter plots to compare ratings with budget.
• Highlight trends in average ratings over time using line plots.
• Annotate key points or outliers.
Git
As we learn, we need to have some way to save our code and projects, and to share it
with others once we build some portfolio-worthy stuff. To do that, we’ll use Git and
GitHub.
Git is a tool for collaboration and keeping track of your work. Whether you’re working
on a solo project or as part of a team, Git helps you save, organize, and track changes
in your codebase. It also makes it easy to experiment without fear of losing work, and
to share your progress with others. GitHub is the UI for us to interact with Git.
What To Know

• Repositories (Repos): These are like folders for your project that track all changes to
les.
• Version Control: Learn how Git tracks changes to your les and allows you to move
back to previous versions if needed.
• Branches: Use branches to experiment with new features or ideas without
affecting your main codebase.
• Merging: Combine branches back into the main branch once your changes are
ready.
• Con ict Resolution: Handle situations where multiple people make changes to the
same part of a le.
• How to use the GitHub platform to interact with Git.

Basic Commands to Master

git init - Create a new repository.
git add - Stage changes for a commit.
git commit - Save a snapshot of your project.
git push and git pull - Share changes with others or pull updates from a shared
repo.
git clone - Copy an existing repository.
git branch and git merge - Create and merge branches.
Learning Resources
• Learn Git Branching
• Game to learn (Oh My Git)
• Data Camp
• CodeAcademy
• W3 Schools

Timeline
Spend ½ week learning the basics of Git.

Example Projects
Start by creating a personal project repository and practicing
commands. As you work on projects, use Git consistently to develop
good habits.
Check-in
At this point, you’ve been studying for about 2.5 months. If you started in January,
we’re now around mid-March, and you’ve already built a strong foundation in Python,
Statistics, SQL, and Data Visualization. Here’s the neat part: You now have the skills to
be a Data Analyst.
This doesn’t mean you’ve completed your journey to becoming a Data Scientist, but it
does mean you’re ready to start marketing yourself as someone who’s job-ready now.
So at this point we’re going to take some time to start putting yourself out there in
the job market.

1. Update Your Resume and LinkedIn

• Add all the skills you’ve learned to your skills sections.
• Add a section for your projects, brie y describing what you did, the tools you used,
and the outcomes.
• Make sure your LinkedIn photo is professional and approachable.
• Update your resume summary and LinkedIn headline/description to use keywords
from jobs you’d like to target. Market yourself as already being the thing you want
to be. So, your headline shouldn't say, “Aspiring Data Analyst,” just Data Analyst.

2. Update GitHub
Make sure your GitHub is super clean with all your work so far (we’ll talk more about
this in the Marketing Your Experience section).

3. Start Applying to Jobs

While it may seem a little early, we’re going to start applying to jobs now.
The idea here isn’t really to get a job now. If that happens, that's awesome, but the
idea is more to start getting a feel for what job descriptions look like, the skills they’re
looking for, keywords to target, and how to tailor your resume and cover letter to
these roles.
Applications at this point are more information-gathering than anything.
Now, back to studying.
Math for Machine
Learning
Math is often the part of data science that intimidates people the
most. The truth is, you don’t need an advanced understanding of
calculus or linear algebra to get started in the eld. While a deeper
grasp of these topics is useful for understanding machine learning
algorithms and optimization techniques, it’s not critical to be an
effective practitioner in your early career. Many Data Scientists –
myself included – developed a more in-depth understanding of
math on the job as their work demanded it.
That said, having some basic math fundamentals under your belt
will make it easier to understand key concepts like gradient descent,
optimization, and machine learning in general.
Linear Algebra is important for understanding how data is
represented and manipulated in machine learning models (think
vectors and matrices). To get started, we need to understand
vectors, matrices, matrix operations (addition, multiplication), and
the concept of dot products.
Calculus plays a role in optimization, such as understanding how
gradient descent works to minimize error in machine learning
algorithms. At the beginning, focus on derivatives and partial
derivatives—enough to understand how gradient descent works—
and learn the concept of gradients and optimization.
We are not going to spend time understanding how to calculate this
stuff by hand. Instead, we’re just going to get an intuition for what’s
going on.

Learning Resources
Luckily, the perfect resource for this already exists: Just watch
3blue1brown’s Essence of Linear Algebra and Essence of Calculus
playlists.
If you want to dig deeper after that, I’d recommend:

• Deeplearning.AI Coursera Specialization (af liate link)

• Imperial College London Coursera Specialization (af liate link)
• Khan Academy: A classic resource for understanding the basics of
calculus and linear algebra in a beginner-friendly format.
• Math for ML Book

Timeline
Plan to dedicate two weeks to learning these math fundamentals.
Again, focus on understanding the intuition of what is happening vs.
the calculations themselves. You can always come back to the math
later in your career.
Machine Learning
Now we’re ready to get started with the most fun part, in my opinion. Machine
learning!
At its core, data science is about solving problems using data. Machine learning is one
of the most powerful tools in your toolkit because it enables you to uncover patterns
and make predictions from data.
Before diving into the speci c algorithms, there’s a lot we need to understand rst:

• Data Cleaning: Learn how to x messy data, handle missing values, and deal with
inconsistent entries.
• Feature Engineering: Turn raw data into useful inputs for your model by creating
features.
• Bootstrapping: Understand how to use sampling to estimate the properties of
your data.
• Feature Preparation: Handle common issues like scaling, imputing missing values
(NULLs), and normalizing your features.
• Train-Test Split and Avoiding Data Leakage: Keep your training and testing data
separate to ensure your models generalize well.
• Exploratory Data Analysis (EDA): Know how to summarize, visualize, and explore
data to spot trends and outliers.
• Supervised vs. Unsupervised Learning: Learn the difference between working
with labeled data vs. nding patterns in unlabeled data.
• Classi cation vs. Regression: Understand when to predict categories
(classi cation) vs. continuous numbers (regression).
• Metrics: Master regression metrics (e.g., MAE, RMSE) and classi cation metrics
(e.g., accuracy, precision, recall, F1, and AUC score).

Once you have some of this background, you can get started with some of the
fundamental algorithms:

• Linear regression
• Logistic regression
• Decision Trees
• K-means Clustering
Learning Resources
• Andrew Ng's classic ML Specialization on Coursera (af liate link)
• A Gentle Introduction to Machine Learning StatQuest Playlist
• DataCamp Machine Learning Fundamentals in Python
• Google's Machine Learning Crash Course

Timeline
Expect to spend about four weeks learning the fundamentals.

Example Projects
Work through all of the projects in the courses.
 • Over tting: Understand what
it is and how to tackle it.
• Regularization: Learn
techniques like ridge and lasso
regression to keep models
from over tting.
• Cross-Validation: Validate your
models properly by splitting
data into multiple
training/testing sets.
• Principal Component Analysis
(PCA): Use PCA to reduce
dimensionality and focus on
the most informative aspects
Then, we’re ready to move of your data.
on to some more • Hyperparameter Tuning:
advanced concepts to Learn how tweaking model
help us handle real-world settings can boost
challenges. performance.
• Ensemble Models: Get familiar
with random forests and
gradient boosting models (e.g.,
XGBoost).
• Time Series Basics: Learn how
to analyze data over time,
focusing on trends, seasonality,
and simple forecasting
methods.
• Introduction to Deep
Learning: Understand the
basics of neural networks, how
they work, and when to use
them.
Learning Resources
• Time Series models
• Gradient Boosting
• Deep Learning:
◦ 3Blue1Brown Learning Intro
◦ Andrew Ng’s Deep Learning Specialization (af liate link)
• Bonus for ML specialists: Designing Machine Learning Systems
(af liate link)

Timeline
Expect to spend about four more weeks getting familiar with these
concepts. You won’t be an expert by any means, but this is kind of a
survey of what exists right now that you could use as tools in future
projects.

Example Projects
To put your learning into practice, do all of the exercises in the
courses. We’ll turn to self-directed projects soon.
Cloud Platforms
Speaking of things you should be familiar with but de nitely won’t
have time to master, we need to learn a little bit about cloud
platforms.
Cloud platforms like AWS, Google Cloud Platform (GCP), and
Microsoft Azure are essential tools for modern data science. They
allow you to work with large-scale data, deploy machine learning
models, and manage resources ef ciently. While you won’t need to
become an expert in all of them, you do need to know what they do
and how to use key features in your work ow.
AWS is the most popular, so unless you already know you’re
targeting a company that uses GCP or Azure, just focus on AWS. The
concepts are similar with the others and you can always learn them
later.
What to Know
• Storage and Databases: Learn how cloud platforms handle data
storage (e.g., Amazon S3) and how to access data securely.
• Compute Services: Understand how to use cloud resources to run
large-scale computations or train models (e.g., AWS EC2).
• ML Services: Get familiar with cloud-based tools like AWS
SageMaker for training, deploying and managing machine
learning work ows.
Learning Resources
• Fully-Managed Notebook Instances with Amazon SageMaker - a Deep Dive
SageMaker playlist
• AWS SageMaker For ML And DL Tutorial Playlist

Timeline
• Spend one week exploring the basics. Start with free-tier accounts on AWS, GCP,
or Azure to get hands-on experience.
• Dive deeper during projects when you need to deploy models or work with cloud-
based data storage and analytics.

Example Projects
Walk through example tutorials in your own environment to get started. Then, use
cloud resources in your self-directed end-to-end projects (next learning phase!)
Business Logic
Having strong business intuition is also crucial for Data Scientists. It’s
not enough to build models—you need to work on projects and
interpret results in ways that drive business impact. This is usually
tested via case studies, which are designed to test your ability to
think critically about business problems, identify key metrics, and
propose actionable solutions. These skills are essential for success in
interviews, and obviously on the job where business impact is
ultimately the thing that will drive your career forward.
What to Know
1. Understand common causes for changes in metrics, such as
seasonal trends, product changes, or external factors.
2. Break down metrics into components (e.g., customer acquisition
vs. retention) to pinpoint areas of change.
3. Key Metrics for Different Business Types:
3.1. E-commerce: Conversion rates, average order value (AOV),
customer lifetime value (CLV).
3.2. SaaS (Software as a Service): Monthly recurring revenue
(MRR), churn rate, user engagement.
3.3. Marketing: Return on investment (ROI), click-through rate
(CTR), cost per acquisition (CPA).
3.4. Operations: Ef ciency metrics, downtime, capacity utilization.
4. How to Answer Case Studies. We’ll talk about this in the interview
section.
Learning Resources
• YouTube mock interviews
• InterviewQuery has practice questions where you can read
others’ answers.

Timeline
• Spend ½ week understanding the basics of business metrics and
structured problem-solving. Basically, just get your mind thinking
about these kinds of things in general. It will take a lot longer to
get good at business intuition and answering case study
questions.
• Dedicate additional time to practice and preparation before each
interview. Regularly work on mock case studies to re ne your
approach and gain con dence.
LeetCode
The last thing on our roadmap is “LeetCode-style" coding.
If you’re not sure what that means, basically it’s a type of coding
question where you’re given algorithmic challenges to solve. These
challenges test your understanding of data structures (like arrays,
linked lists, and trees), algorithms (such as sorting, searching, and
dynamic programming), and computer science fundamentals.
I personally despise LeetCode for data science interviews, because
in my opinion it’s completely irrelevant for the kind of coding we do
on the job. Data science typically involves wrangling messy data,
building predictive models, and implementing data pipelines — not
designing the most optimal binary search algorithm. But the fact is,
a lot of companies still use it, so you need to know how to pass the
interview.
That being said, for most data science interviews it’ll just be
questions on arrays and strings, so we don’t need to go into software
engineer-level learning on data structures and algorithms.
What I’d suggest learning is:

• Basic familiarity with Big-O notation.

• Common edge cases and how to approach coding problems in
general by asking clarifying questions.
• String and Array questions.
• Practice coding and speaking at the same time (not easy! You
need to really practice this).

Also, quick note that data science interviews also often test SQL, so
don’t forget to practice that as well.
Learning Resources
• Big-O Notation
• NeetCode
• Grokking the Coding Interview Patterns in Python
• You can also just walk through the LeetCode arrays questions
and ask ChatGPT to help you understand the “tricks” and optimal
solutions.

Timeline
Dedicate about ½ week to familiarizing yourself with LeetCode and
solving introductory problems. Then, start a habit of practicing one
Python and one SQL question every day until you get a job.
Summary of Skills to Learn
Primary Topic Speci c Skills Time to Learn
Python - Basics: Variables, data types, 4 weeks
loops, conditionals, functions
- Data structures: Lists,
dictionaries, sets, tuples
- File handling, exceptions
- Libraries: Pandas, Numpy,
Scikit-learn
- Package management: pip,
conda
Statistics - Descriptive statistics: Mean, 4 weeks
median, variance, distributions
- Inferential statistics:
Hypothesis testing, p-values,
con dence intervals
- Experimentation: A/B testing,
sample size, randomization
- Probability: Bayes' theorem,
basic concepts
- Multicollinearity, correlation,
causation
SQL - Core querying: SELECT, 2 weeks
WHERE, ORDER BY, LIMIT,
NULL handling
- Joins: INNER, LEFT, RIGHT,
FULL OUTER
- Aggregations: GROUP BY,
COUNT, AVG, SUM
- CASE statements, CTEs,
window functions, subqueries
Summary of Skills to
Learn
Primary Topic Speci c Skills Time to Learn
Data Visualization - Python: Matplotlib, Seaborn, 1 week
Plotly
- Dashboards: Tableau, Power BI
Git - Basic commands: init, add, 0.5 weeks
commit, push, pull, clone
- Branching and merging
- GitHub: Repository setup,
documentation
Math for ML - Vectors, matrices, dot products, 2 weeks
matrix operations
- Derivatives, partial derivatives,
gradients and gradient descent
Machine Learning - Fundamentals: Regression, 8 weeks
classi cation,
supervised/unsupervised learning
- EDA, data cleaning, feature
engineering, train-test split
- Metrics: RMSE, F1, accuracy, recall,
precision
- Fundamental aglorithms:
Regression, decision trees, k-
means
- Regularization, cross-validation,
hyperparameter tuning
- Ensemble models, time series,
deep learning
Summary of Skills to
Learn
Primary Topic Speci c Skills Time to Learn
Cloud Platforms - Storage: e.g. S3 1 week
- Compute: e.g. EC2
- ML Services: e.g.
SageMaker
Business Logic - Metric analysis: 0.5 weeks
KPIs, A/B testing
- Industry-speci c
metrics: SaaS, e-
commerce,
marketing
LeetCode Practice - Arrays, strings, Big- 0.5 weeks and
O, edge cases ongoing
- SQL challenges
Extra Credit
There are some more things that I think will be really valuable to you
in your rst job, but aren’t essential. If you have time and curiosity, I
would also recommend learning about:

• How to test your code.

• Object-Oriented Programming.
• Code Reviews: Understand how to give and receive constructive
feedback on code.
• Docker & Packaging: Learn the basics of containerization and
how to make your projects portable.
• Air ow: For building data pipelines.
If you read these two books, you'd be off to a very solid start on the
above (and more!):
• Designing Machine Learning Systems (af liate link)
• Software Engineering for Data Scientists (af liate link)
Things NOT to Learn
We’ve covered A LOT. But there is so much more! Here are things I
would skip for now:
• Database Creation and • Kubernetes: While container
Optimization (e.g. indexing, orchestration is essential for
sort keys, etc.). This is large-scale deployment, most
absolutely good to know, but entry-level Data Scientist roles
not critical for entry-level Data don’t involve heavy
Scientists). deployment responsibilities.
• Deep Learning for Niche • NoSQL Databases: Databases
Applications: Working with like MongoDB, Cassandra,
GANs, transformers, or and DynamoDB are good to
advanced architectures is know but less critical than
typically more relevant for mastering SQL for relational
specialized roles in computer databases.
vision or NLP. • Graph Databases: Neo4j and
• Reinforcement Learning: This similar tools are great for
is highly specialized and rarely analyzing network data but
required for most data rarely required in typical Data
science generalist roles. Scientist roles.
• Hadoop & Apache Spark: • Edge Computing and IoT
While useful for large-scale Analytics.
data processing, most Data • Nitty gritty on how LLMs
Scientist roles don’t require work. It's a good idea to get
this, and there are typically some experience with
software engineers to help (or prompting and working with
simpler services to get the job APIs, but you’re unlikely to be
done). asked to code a Transformer
• Kafka: Real-time data from scratch in an entry-level
streaming tools are more data science generalist role.
relevant for engineering-
focused roles.
How to Get
Experience and Build
a Portfolio
By now, you’ve gained foundational skills in Python, statistics, SQL,
machine learning, and more, and have completed a few small
follow-along projects. But here’s the next challenge: Hiring
managers want to see real-world experience. How are you going to
get experience, if no one will hire you without experience?
What we’re going to do is create our own opportunities. Here’s
how to do it step-by-step.
Building a Portfolio
Align Your Projects with Your
1.
Career Goals
At this point you should be pretty familiar with the kinds of jobs that
are out there, since you’ve been casually applying for a few months.
Align your project plan with the expectations of the industries you’re
targeting. For example:

• E-commerce: Recommendation systems, customer

segmentation, or sales forecasting.
• Finance: Fraud detection, credit risk modeling, or portfolio
optimization.
• Healthcare: Predictive analytics for patient outcomes or medical
image analysis.
• Tech: Sentiment analysis of social media, churn prediction, or A/B
test analysis.

Also, consider the kind of role you’re aiming for:

• Analytics roles: Highlight SQL, dashboards, and data visualization.

• Machine learning roles: Focus on building and deploying
predictive models end-to-end.
• Generalist roles: Showcase versatility with a mix of data
wrangling, analysis, and modeling.
2. Choose Projects That Showcase
a Full Workflow
To stand out, your portfolio projects should demonstrate an end-to-
end data science work ow. That means showing your ability to:
• De ne the problem: Start with a clear, actionable question or
goal.
• Collect and clean data: Source raw data through web scraping,
APIs, or public datasets. This is super important! We don’t want to
have a project with an unrealistically-clean dataset. It should be
something a little more messy and representative of the real-
world.
• Perform Exploratory Data Analysis (EDA): Use visualizations and
summary statistics to uncover patterns and outliers.
• Engineer features: Create meaningful inputs for your models
through scaling, encoding, combining data, and handling
missing values.
• Build and evaluate models: Train, test, and validate models with
clear metrics. Highlight how your model solves the problem
effectively.
• Communicate insights: Translate ndings into actionable
recommendations, using visualizations or dashboards. Consider
using Tableau Public, Power BI, or Streamlit to create interactive
visuals and shareable dashboards.
• Optionally, deploy the model!
3. Pick a Project Idea
Now that you know all the pieces that should go into a project, it’s time to pick a
speci c project to work on. Ideally, pick something that is actually interesting to you.
You’re going to be thinking about it for weeks, and you’re going to do a better job if
it’s fun and showcases your personality.
If you’re not sure where to start, here are some project ideas:
Public Data Analysis

• Social Media: Use Twitter, Reddit, or YouTube APIs to analyze trends, sentiment, or
user behavior.
• Economic Data: Analyze public datasets like World Bank or Census data to identify
trends in GDP, unemployment, or urbanization.
• Weather Data: Use NOAA or OpenWeatherMap APIs to analyze patterns and build
forecasting models.

Recommendation Systems

• Collaborative Filtering: Build a movie, book, or product recommendation system

using user-item interactions.
• Content-Based: Recommend items based on attributes like product features or
user preferences.

Sentiment Analysis

• Text Analysis: Analyze customer reviews, social media posts, or product feedback
to assess sentiment or extract insights.
• Topic Modeling: Use NLP techniques to identify themes in large text datasets.

Predictive Models

• Churn Prediction: Predict customer churn for a subscription service or product.

• Fraud Detection: Identify fraudulent transactions in nancial data.
• Time-Series Analysis: Build forecasting models for stock prices, sales data, or
environmental variables.
4. Present Your Project
Once the project is complete, ensure it’s presented in a professional
and engaging way.
It’s up to you whether you’d rather have a separate professional
website for your portfolio, or just make your GitHub look really nice.
Also, every time you nish a project, make sure to add it to your
resume and LinkedIn. In the next section on marketing your skills I’ll
go way more into detail about this.
Now you have demonstrated you can actually do some stuff! It’s
time to use those skills to get real-world experience.
Real-World Project
My favorite “hack” for getting real-world experience is to offer your
services for free to real clients.
I realize this is controversial. Yes, in an ideal world we'd all be paid for
our time, but the fact is that as a beginner your options are limited.
It's worth it to invest your time for a big payoff once you transition
into the eld.
These projects are valuable because they allow you to work with
actual data, address real business problems, and showcase your
ability to apply data science techniques in practical scenarios. Here’s
how you can take the initiative to gain this experience:
Start Where You Are
If you’re currently employed, look for ways to apply data science
within your organization, even if it’s outside your of cial role. For
example:
• If you work in retail, analyze sales trends to optimize inventory
management.
• In a small of ce setting, automate repetitive tasks or analyze
performance metrics.
• If you’re in customer service, explore feedback data to identify
trends or improve satisfaction scores.

These internal projects not only add to your portfolio but also
demonstrate initiative, problem-solving skills, and an ability to add
value in a business context. Often, success in these projects can lead
to formal recognition, such as transitioning into a more analytical
role within the same company. Just make sure you’re following all
the rules about handling this data.
Leverage Your Network
If you aren’t currently employed or your current job doesn’t provide
opportunities for data science work, there are still plenty of options.
Start with people you already know—friends, family, or
acquaintances—who may own businesses or manage teams. For
instance:

• Help a friend’s family-owned restaurant analyze customer reviews

to improve service.
• Build a pricing optimization model for someone running an e-
commerce shop.
• Develop a tool to track and predict seasonal trends for a local
business.
Pitch to Small Businesses or Nonpro ts
If you don’t know anyone, reach out to local businesses and
nonpro ts.
Small businesses and nonpro ts often lack the resources for
advanced analytics but can bene t signi cantly from your skills.
Reach out to these organizations with a speci c, actionable
proposal. For instance:

• Offer to create a dashboard to track donation trends for a

nonpro t.
• Analyze sales and customer data for a local store to help optimize
marketing efforts.
• Develop a basic predictive model for a startup to improve
customer retention.
When pitching, focus on the business impact of your work. Most
small organizations will not fully understand data science concepts,
so frame your proposal in terms of how it will save time, reduce
costs, or improve ef ciency. Even if these projects start as unpaid,
they often lead to recommendations, paid opportunities, or at the
very least, excellent additions to your portfolio and resume.
Volunteer with Existing Organizations
Organizations like DataKind are excellent ways to gain real-world
experience while contributing to meaningful causes. These
platforms connect Data Scientists with nonpro ts in need of
expertise. Projects range in scope from guring out survey design to
developing computer vision models.
Not only do these experiences enhance your technical skills, but
they also allow you to learn from more experienced Data Scientists
and grow your network.
Keep Building Independently
While real-world projects with clients or businesses are ideal, don’t
let the lack of immediate opportunities stall your progress. If
responses from potential clients are slow, keep developing portfolio
projects on your own.
The goal is to maintain momentum, re ne your skills, and add depth
to your portfolio while waiting for opportunities to come through.
Marketing
Your
Experience
If you've been following the year-long timeline, but now it’s around
mid-August. You’ve learned the basics and even put it to use on
some self-directed and/or real-world projects. Now it’s time to focus
on updating your application materials (again) to re ect your new
skills and position yourself as a strong candidate in the data science
eld.
Portfolio
First, your portfolio. I mentioned before that you can choose to have a website or just
a nicely-formatted GitHub. Personally, I’d just go with Github. Here are some
awesome examples:

• https://github.com/katiehuangx
• https://github.com/TeneikaAskew
• https://github.com/Erlemar

These all have a clear and engaging introduction, and easily direct the reader to
projects.
If you’ve set up a professional website, make sure it includes:

• A brief but engaging bio

• Links to your GitHub and LinkedIn
• Clear sections highlighting your projects with descriptions of the problem, your
approach, and results
• Contact information

Regardless of which approach you take, each project should have its own clean,
organized GitHub repository for each project, including:

• A README with the project overview, tools used, and key ndings.
• Well-commented code and modular scripts.
• Instructions for reproducing the project (if applicable).

One thing to highlight in your write-ups is why you made the decisions you did
during the project. For example, why did you choose a speci c model or pre-
processing approach? This will help prepare you for interviews and show that you
understood the concepts well enough to make good decisions.
It’s also useful to write about your project on Medium or LinkedIn to explain your
process and results. You’ll want to focus on storytelling: what problem you solved,
how you solved it, and why it matters. Remember, your communication skills are also
important for data science, so creating an engaging narrative also highlights an
important skill for future employers.
LinkedIn
Next, we’re going to make sure your LinkedIn looks good. We’ve already done some of
this in the mid-learning check-in, but here’s a checklist to make sure everything is
polished and optimized:

✓ Custom URL: Create a clean and professional LinkedIn URL (e.g.,

linkedin.com/in/yourname).
✓ Professional Photo: Use a high-quality photo with a friendly and approachable
expression.
✓ Headline: Present yourself as already being in the eld, such as “Data Scientist |
Machine Learning Enthusiast,” rather than “Aspiring” or “Student.”
✓ Skills: Add a comprehensive list of data science and related skills. Make sure to
update this with things you’ve learned since last time, and keywords you’ve
noticed in relevant job postings.
✓ Projects: List key projects with concise descriptions and links to your portfolio or
GitHub.
✓ Pro le Summary: Use keywords and action-oriented language to describe your
expertise, skills, and career aspirations.
✓ Links: Include links to your GitHub, portfolio website, and relevant articles or blog
posts.
✓ Desired Roles: Add titles and industries you want to be contacted for, such as “Data
Scientist,” “Machine Learning Engineer,” or “Data Analyst.”
✓ Experience Titles: If you’ve done unpaid or freelance work, list those projects as
legitimate roles. For example, label yourself as a “Data Scientist” even if it was an
unpaid gig.
✓ Stay active: Share relevant projects, articles, or insights regularly to keep your
pro le active and increase visibility.

Some good LinkedIn examples:

• https://www.linkedin.com/in/katiehuangx/
• https://www.linkedin.com/in/teneikaaskew/
Resume
When updating your resume, keep it concise and tailored:

✓ Place relevant technical skills at the top.

✓ Highlight your projects, especially real-world ones, with speci c
details about your contributions, tools used, and measurable
outcomes.
✓ If you worked on a freelance or volunteer project, list your title as
“Data Scientist” or “Data Analyst” (just like we did on LinkedIn)
✓ Study job descriptions for roles you want and ensure your resume
includes relevant keywords and phrases.
✓ Add links to your GitHub, portfolio, and LinkedIn.
✓ Present unrelated experience in terms of transferable skills—e.g.,
problem-solving, teamwork, or leadership.

A strong beginner data science resume should have:

✓ One page only

✓ A clear summary or heading at the top that aligns with your
career goals (e.g., “Data Scientist | Machine Learning Engineer”)
✓ Real-world projects and volunteer work
✓ Skills, tools, and technologies used in projects listed in your
experience
✓ Links to your portfolio, GitHub, and LinkedIn

This advice is somewhat speci c to the U.S., but the core ideas
remain the same even if you live in a country that expects longer
CVs.
Let’s look at an example resume.
 • It's one page and formatted in
an ATS-friendly way.
• Links are easy to nd.
• The summary has lots of
keywords, and describes the
candidate as already being in
the eld.
• Skills are prominently listed at
the top.
Here you can see: • Since there is a relevant
certi cate (or degree),
Education is near the top as
well.
• Then projects follow, with clear
examples of tools used and
metrics.
• Finally, professional experience
is at the end, where we tried as
much as possible to make it
relevant.
Cover Letters
Now, cover letters: While I don't personally rely on cover letters too
heavily at this point in my career, I think in the early days having a
well-crafted cover letter that shows enthusiasm for this particular
role at this particular company (with details), and transferable skills
can go a long way.
For example, assuming we’re the teacher with the resume above,
and we want to apply for a Junior Data Scientist at a bank:
How to Get Interviews
Now that we have our application materials ready and a strong
portfolio, it’s time to start applying in earnest.
We often hear that applications are just a numbers game. And while
that is true to an extent, we need to be a lot more strategic than just
sending our resume out to every posting we see (and we absolutely
don’t want to be doing this with GenAI tools).
Instead, prioritize jobs where:
1. You have a strong chance, or
2. You would really love the role.
Spend time customizing your resume – or better yet, developing a
networking strategy, which we’ll talk about next – for those roles.
If you still have time after really focusing on the high probability or
high payoff positions, then sure, go ahead and spam your resume a
bit.
What we want to do is avoid the cycle of mindlessly sending out
applications, getting rejected, feeling like there’s no point putting in
the effort, so we’re lazy with the next batch of applications, and so
on. This is a really poor approach strategically, and is bad for your
mental health.
Networking
Even more promising than applications is networking and putting
yourself out there one-on-one. In my data science mentoring, the
people I’m seeing be successful are the ones with very little social
fear. They are super proactive about reaching out to recruiters, hiring
managers, and in-person networking.
I know this sounds painful – trust me, I hate doing it too – but the
fact is that this is absolutely the highest leverage tool we have in our
toolbelt.
If you have little or no industry experience, automated resume
screening systems might lter out your application. Reaching out
directly to a recruiter can bypass this initial screen and give you a
chance to sell yourself as a candidate.
Recruiters receive a lot of messages, so keep yours short, respectful,
and to the point.
Start by brie y introducing yourself, mentioning your relevant skills,
projects, or certi cations. Explain why you’re interested in the
company speci cally, rather than sending a generic message.
For example:
“Hi [Recruiter’s Name],
I’m [Your Name], a Data Scientist with skills in Python, SQL, and machine learning. I
recently completed a project where I built a recommendation system to improve
customer engagement for an e-commerce dataset. I’m particularly interested in
[Company Name] because of your focus on [speci c product, value, or challenge]. I’d
love to connect to learn more about what you look for in data science candidates.
Thank you for your time!”
Recruiters and hiring managers are looking for people who can solve real problems.
Instead of just stating you’re interested in the role, we can even take it a step
further by showing how your skills and projects could add value.
Study the company’s products, recent news, and any challenges they might be
facing. LinkedIn, company websites, press releases, or even industry reports can give
you ideas. In your message, suggest a few ways your data science or ML skills could
address speci c needs. For example, if the company focuses on e-commerce,
mention how your experience with recommendation systems or customer
segmentation projects could enhance their user experience.
Even if you don’t have direct industry experience, reference any relevant projects,
volunteer work, or skills that demonstrate your ability to solve similar problems. For
instance, mention a project where you analyzed customer data or a volunteer role
where you optimized a process – anything that shows how you’d contribute
effectively.
Whether you’re applying to job postings or reaching out to potential contacts, one
super important thing is to remember that you’re not going to get a job by focusing
on how great the job would be for you, but how you’re the best person to solve the
company’s problems. Make sure to clarify what differentiates you from other
applicants, and how these speci c skills are best positioned to solve the speci c
problems outlined in the job description. You need to understand the role, and
communicate clearly how you plan to bring value to the company – not the other
way around.
For example:
Instead of saying: “This role is exciting to me because I’m passionate about machine
learning.”
Say: “I see this role focuses on customer segmentation, and my recent project using
clustering algorithms demonstrated how to drive actionable insights from user data.”
Be consistent about reaching out to potential connections and being thorough with
your applications for high-potential jobs, and at one point you’ll start being called for
interviews.
Interview Prep
Once you get an interview on the calendar, it’s time to start
preparing in earnest.
The key thing that I do differently with interview prep is,
essentially, over-prepare. I put in a ton of work in advance to
anticipate what might be covered, and dig deep into studying those
topics so that it looks like I came up with clever, well thought-out
solutions on the y.
This is particularly important if you lack signi cant industry
experience. Preparation will help you bridge that gap and present
yourself as someone who’s ready to learn and contribute from day
one. You cannot afford to go into an interview unprepared.
Interviews can vary widely depending on the company and the role,
so focus on the common components:
Recruiter Screen
Be prepared to give a brief introduction, summarize your background, and explain
why you’re interested in the role.
At this stage, it’s important to make a strong case for why your skills align with the
job requirements. Use keywords from the job description, reference relevant details
from their blog or website, and ensure your elevator pitch is well-rehearsed.
Review your past projects thoroughly, understanding every detail, including the
reasoning behind your decisions. This preparation ensures you’re ready if the recruiter
asks you to walk through a speci c project. Whenever possible, tailor your examples
and explanations to align with the requirements for the job.
Finally, ask thoughtful questions to gain clarity about the role and understand the
next steps in the interview process. This not only shows your interest but also helps
you prepare effectively for future stages.
Coding Questions
By now, you’ve been practicing coding, so you should be prepared.
Make sure to practice thinking out loud and coding at the same
time.
Remember, it’s not just about getting the correct answer —
interviewers are evaluating how you think, approach problems,
ask clarifying questions, and collaborate. It’s worth taking the time
to practice coding interviews with another person who can provide
feedback on these aspects. This will help you re ne your
communication skills and improve your ability to explain your
thought process effectively.
Behavioral Questions
You can get hints about the kinds of behavioral questions you might
be asked by reviewing the company’s values or leadership
principles. These often guide the questions interviewers ask to
evaluate cultural t and alignment with their priorities. Prepare at
least one example from your past experience that demonstrates
how you embody each of these values.
When answering, use the STARR method (Situation, Task, Action,
Result, Re ection) to structure your responses effectively. This
ensures your answers are clear, concise, and highlight your impact.
Focus on examples that showcase teamwork, problem-solving, and
adaptability while tying them back to the company’s core values.
Case Studies
Case studies test your ability to solve real-world problems using data
science and machine learning. Prepare by researching the business
you’re interviewing with, understanding their industry, and
anticipating relevant problems, metrics, and data types. Dive into
the company’s products, services, and goals to tailor your approach
and demonstrate domain-speci c knowledge.
Business
Context
Be ready to discuss how your
work as a Data Scientist can
impact the company’s bottom
line, decision-making processes,
or strategic goals. This shows that
you not only have technical
expertise but also understand the
business implications of your
work. Tailor your responses to
align with the company’s mission
and priorities.

Prepare
Questions for
Each Phase
For every interview stage, have
thoughtful questions prepared.
These could be about the role, the
team, the company’s challenges,
or their expectations. Asking
questions not only shows your
genuine interest but also helps
you gain insights to better
position yourself as the ideal
candidate.
Preparation is Key, But You Donʼt
Need to Know Everything
The key to success in all of these is to think in advance about the
business to such an extent that you’re able to essentially guess what
the case studies will be, so that you can narrow down the areas to
study.
I have a detailed video that goes over how to prepare for every phase
of the interview process, from the recruiter call to behavioral
questions to case studies. I would de nitely recommend checking
that out!
All of the advice in that video is the same for beginners, except that
you will have more times when you do not have relevant
experience to draw on. In those cases, don’t panic—pivot the
conversation to how you would approach the problem. For example:
“While I haven’t worked on a problem like this before, here’s how I
would break it down…”
This demonstrates critical thinking and a willingness to tackle new
challenges, both of which are valuable traits for any candidate.
It’s also ok to ask clarifying questions of the interviewer and
demonstrate your curiosity and growth mindset. Interviewers don’t
expect you to know everything, but they do want to see how you
think and approach complex problems.
Remember, every interview you attend—even if it doesn’t result in
an offer—is a valuable learning experience. The more you practice,
the more comfortable and con dent you’ll become.
How to Do Well When
You Start Your First
Job
Congratulations—you’ve landed your rst data science role! Starting
a new job can be intimidating, especially in a technical eld like this
one.
Here are some suggestions on how to do well when you start your
rst job:

1. Understand the Business Context

Your rst priority should be understanding the company’s goals and
how your work ts into the bigger picture. This means:

• Learn the metrics that matter: Find out what KPIs (Key
Performance Indicators) your team or company cares about and
why.
• Understand the challenges: Read up on industry trends, internal
documentation, or past projects to get a sense of the business’s
pain points.
• Talk to stakeholders: Ask stakeholders or team members what
problems they’re trying to solve and how your work can help.
Building this understanding early on will guide your decisions
and make your contributions more impactful.
2.Ask Questions and Seek
Feedback
It’s okay not to know everything—no one expects you to at this
point. What matters is that you’re proactive about learning. Here’s
how to approach this:
• Ask questions: Whether it’s about the company’s processes, tools,
or expectations, don’t hesitate to seek clari cation. This shows
initiative and a desire to get things right.
• Get feedback often: Ask for feedback on your work from peers,
mentors, or managers. The sooner you know what you’re doing
well (and what you need to improve), the faster you’ll grow.

3. Leverage Your Team

You’re not in this alone. Your team is your greatest resource, so don’t
hesitate to:

• Pair program with colleagues: This can help you learn coding
practices, tools, and work ows speci c to your company.
• Learn from code reviews: Take reviews seriously and see them as
opportunities to improve your skills.
• Find a mentor: Identify someone on your team who is
experienced and approachable, and learn from their guidance.

Building strong relationships with your colleagues will not only help
you learn faster but also create a supportive environment where you
can thrive.
4. Focus on Delivering Value
In your rst few months, aim to contribute to projects that provide
visible value, even if they’re small. For example:

• Automate a repetitive task.

• Create a simple dashboard or visualization to help stakeholders
make decisions.
• Optimize an existing process or improve a basic model.
Delivering results, even incremental ones, builds trust and
establishes your reputation as someone who gets things done.

5. Stay Curious and Keep Learning

The learning doesn’t stop once you’re hired (or ever, truth be told).
Continue to build your skills by:
• Exploring new tools and techniques: You’ll likely encounter tools
or methods you haven’t used before—take the time to master
them, even outside of work.
• Learning from mistakes: If something doesn’t work, treat it as a
learning opportunity.
Final Thoughts
Becoming a Data Scientist in 2025 might feel like a monumental
challenge, but with focus, persistence, and a clear roadmap, it’s
absolutely within your reach. Remember, this journey is a
marathon, not a sprint. It’s not about mastering every tool or
algorithm under the sun—it’s about building a strong foundation,
solving real problems, and demonstrating your ability to drive
business value.
If you’re feeling overwhelmed, don’t worry—that’s normal. Take it
one step at a time, stick to the timeline we’ve laid out, and keep
pushing forward. The most important thing is to stay consistent, stay
curious, and keep building—whether it’s your skills, your portfolio, or
your network.
If you’re worried about systems to put all this into practice, I have a
video on how I study consistently with a full-time job (learning never
ends in this eld) – I’d recommend checking that out next so you
have the best possible chance of success in the coming year.
Thanks for sticking with me to the end of this roadmap, and good
luck! You’ve got this.
 • ”Marina
My YouTube channel:
Wyss - Gratitude Driven
• ÙshareSignweekly
up for my newsletter! I

Connect
ideas and
insights across personal
growth, professional

With Me
development, and the world of
AI and data science.
• ™ If you'd like to chat with me
1:1, you can book a call here:
https://topmate.io/marina_wyss
• ❤ Or, nd me on Instagram:
link