An AI-Driven Resume Analyser with Job

Recommender and Portfolio Code Generator

Mrs. Vidya Kawtikwar Mr. Lobhan Patil Mr. Pratham Mistry
Assistant Professor Student Student
Department of Computer Engineering Department of Computer Engineering Department of Computer Engineering
St. John College of Engineering and St. John College of Engineering and St. John College of Engineering and
Management Management Management
Palghar, India Palghar, India Palghar, India
[email protected] [email protected] [email protected]

Mr. Shreyash Narse Ms. Siddhi Bari

Student Student
Department of Computer Engineering Department of Computer Engineering
St. John College of Engineering and St. John College of Engineering and
Management Management
Palghar, India Palghar, India
[email protected] [email protected]

Abstract—JobHelper is an AI-driven platform designed to ease by selecting from customizable templates designed to
optimize and simplify the job application process. By suit a variety of industries and job roles.
generating professional resumes through customizable
templates, the platform ensures that job seekers effectively
present their skills and experience. In addition to resume II. METHODOLOGY
creation, JobHelper analyzes and matches user profiles with
relevant job opportunities, improving their chances of landing The JobHelper system is an AI-driven platform designed
suitable roles. Furthermore, the platform automatically to assist users in resume optimization, job recommendations,
generates professional portfolio websites by providing HTML, and portfolio generation. The system is divided into four
CSS, and JavaScript code, allowing users to easily showcase
interconnected modules, each addressing a specific
their work. JobHelper's integrated approach aims to enhance
job-seeking success and elevate users' professional presence functionality.

Keywords—React, Javascript, NLP, Pyresparser, Resume The methodology section outlines the comprehensive
Analyzer. approach taken to develop the Resume Analyzer and
Recommender System Using Python, detailing the processes
I. INTRODUCTION
and technologies used at each stage of the system's
development. This section covers the design and
An AI-driven resume Analyzer with a Job Recommender implementation of both the frontend and backend
and Portfolio Code Generator is a platform tailored to components, database management, dataflow, security
revolutionize the job application process by leveraging the measures, server setup, database deployment, application
power of artificial intelligence. This innovative tool deployment, and security measures integration.
addresses the core needs of job seekers, from generating
personalized resumes to offering job recommendations and
The Resume Analysis Module begins by parsing resumes
portfolio creation. The project is divided into three distinct
uploaded in various formats (PDF, DOCX) using OCR and
yet interconnected parts, each aimed at enhancing the user's
NLP libraries like spaCy and NLTK. It extracts key skills,
professional profile and increasing their chances of success
detects grammatical and syntactical errors using tools
in the competitive job market.
like Grammarly API, and evaluates the resume's quality
through a scoring algorithm. The Job Recommendation
JobHelper uses advanced AI algorithms to analyze the Module utilizes the extracted skills to match users with
content of a user's resume, extracting key skills, relevant job profiles from a database populated via APIs
qualifications, and experience. By understanding the user's like Indeed or LinkedIn. A recommendation algorithm, such
professional background, it intelligently recommends job as collaborative filtering or cosine similarity, is employed to
listings that closely match the individual's expertise and suggest jobs based on skill relevance, location, and
career aspirations. experience level. The Resume Generation Module integrates
generative AI models (e.g., GPT-4) to create professional
This feature not only saves time but also ensures that job resumes using customizable templates. Users can edit and
seekers are presented with relevant, high-quality preview resumes in real-time before exporting them in
opportunities that align with their skill sets, thus improving formats like PDF or DOCX. Lastly, the Portfolio Code
the likelihood of securing interviews and job offers. The Generation Module generates personalized portfolio
platform incorporates state-of-the-art generative AI websites by dynamically populating pre-designed HTML,
technology to streamline the resume creation process. This CSS, and JavaScript templates with user data. Users can
JobHelper allows users to craft professional resumes with customize the portfolio's design and layout, with the final
code packaged into a downloadable ZIP file. The system is

XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE

built on a modular architecture, featuring a React.js frontend the-art language model, is used to auto-fill resume sections
for a responsive interface, a Node.js/Django backend for with coherent and professional content based on user input.
data processing, and AI/ML models for NLP and Hugging Face Transformers provides pre-trained models for
recommendation tasks. It is deployed on a cloud platform natural language generation, enabling the system to create
(e.g., AWS or Google Cloud) for scalability and evaluated tailored resumes efficiently. These AI models ensure that the
using metrics such as resume analysis accuracy (precision, generated resumes are high-quality and relevant to the user's
recall, F1-score), job recommendation precision (mean profile.
reciprocal rank, precision@k), and user satisfaction scores.
This comprehensive approach ensures a robust, user- Frontend Development: The frontend of the system is
friendly, and scalable solution for job seekers. built using React.js, a popular JavaScript library for building
user interfaces. React.js enables the creation of a responsive
Resume Parsing and Text Extraction: The Resume and interactive web application, allowing users to upload
Analysis Module relies on tools like PyPDF2 and python- resumes, customize templates, and preview outputs in real-
docx to extract text from resumes uploaded in PDF and time. Its component-based architecture ensures modularity
DOCX formats, respectively. PyPDF2 is a Python library and scalability, making it ideal for a feature-rich application
specifically designed for reading and extracting text from like JobHelper
PDF files, making it ideal for parsing resumes in PDF
format. Similarly, python-docx is used to handle DOCX Backend Development: The backend is developed using
files, enabling the extraction of text, headings, and other Node.js or Django. Node.js is a JavaScript runtime that
relevant content. For resumes with non-selectable text (e.g., enables efficient handling of asynchronous operations,
scanned documents), Tesseract OCR is employed. Tesseract making it suitable for real-time data processing. Django, a
is an open source Optical Character Recognition (OCR) Python-based web framework, is used for its robustness and
engine that converts images of text into machine-readable scalability in handling complex business logic. Both
text, ensuring comprehensive text extraction from all resume frameworks integrate seamlessly with the database and
formats. frontend, ensuring smooth data flow and processing.

Natural Language Processing (NLP): The extracted text is Database Management: The system uses MongoDB or
processed using NLP libraries like spaCy and NLTK. spaCy MySQL for database management. MongoDB, a NoSQL
is a powerful NLP library that provides pre-trained models database, is chosen for its flexibility in handling
for tokenization, part-of-speech tagging, and named entity unstructured data, such as resumes and user profiles.
recognition (NER). It is used to identify skills, job titles, and MySQL, a relational database, is used for structured data
other key entities in the resume. NLTK (Natural Language storage, such as job profiles and user preferences. These
Toolkit) complements spaCy by offering additional databases ensure efficient storage, retrieval, and
functionalities like stemming, lemmatization, and sentiment management of data.
analysis, which help in refining the extracted data. These
tools collectively enable the system to analyze and structure Portfolio Code Generation: The Portfolio Code Generation
the resume content effectively. Module uses Bootstrap for responsive web design and
JavaScript for interactivity. Bootstrap provides pre-designed
Grammatical and Syntactical Error Detection: To ensure templates and components, enabling the creation of visually
the resume is error-free, the system integrates tools like appealing and mobile-friendly portfolio websites. JavaScript
Grammarly API or LanguageTool. Grammarly API is a is used to dynamically populate the templates with user data
widely used tool for detecting grammatical, punctuation, and and add interactive features. The final code is packaged into
spelling errors in text. It provides real-time suggestions for a downloadable ZIP file using JSZip, a JavaScript library for
corrections, enhancing the quality of the resume. creating and managing ZIP files.
Alternatively, LanguageTool, an open-source grammar
checker, is used to identify errors and offer corrections, Deployment: The system is deployed on cloud platforms
ensuring the resume adheres to professional standards. like AWS (Amazon Web Services) or Google Cloud. These
platforms provide scalable and reliable infrastructure for
Job Recommendation Algorithms: The Job hosting the application, ensuring high availability and
Recommendation Module uses machine learning libraries performance. Services like AWS EC2 for virtual servers and
like sci-kit-learn and TensorFlow to implement Google Cloud Storage for data storage are utilized to
recommendation algorithms. scikit-learn provides a range of support the system's requirements.
algorithms, including cosine similarity and collaborative
filtering, which are used to match user skills with job Evaluation Metrics: The system's performance is evaluated
descriptions. TensorFlow, a deep learning framework, is using precision, recall, and F1-score for resume analysis,
employed for more advanced recommendation models, such mean reciprocal rank (MRR) and precision@k for job
as neural network-based approaches. These tools ensure recommendations and user satisfaction scores. These metrics
accurate and personalized job recommendations based on ensure that the system delivers accurate and user-friendly
the user's profile. results, meeting the needs of job seekers effectively.

Generative AI for Resume Generation: The Resume

Generation Module leverages generative AI models like
GPT-4 or Hugging Face Transformers. GPT-4, a state-of-
 III. LITERATURE REVIEW
The literature on resume parsing focuses on methods to
my college project title is "JobHelper: An AI Driven automate the extraction of critical details from resumes,
Resume Analyzer with i Job Recommendor and Portfolio streamlining the recruitment process for organizations.
Code Generator" Traditional approaches often rely on rule-based or pattern-
matching techniques, while more recent advancements
"1)It Provides skills based on the resume ,this will provide leverage natural language processing (NLP) capabilities.
the job recommendations to the user Named entity recognition (NER) and syntactic parsing are
crucial NLP tasks employed for identifying and classifying
2) project also has resume analyser will also take out he key elements such as names, contact details, and skills.
gramatical , and all syntatical errors as well as it will anlyze Additionally, multiprocessing techniques have been
the resume and suggest job profiles for the user explored for parallel processing, significantly reducing
processing time, especially with large volumes of resumes.
Evaluation metrics such as accuracy and precision are used
3) It will Generate professional resumes with customizable
to assess system performance, often with manually
templates using a generative AI- when the user uploads
annotated datasets and user feedback.
resume in any format of the web app,
Automated parsing systems find applications in various
domains, offering, benefits such as cost savings and
4) It will generate and give portfolio code to the user which efficiency improvements in candidate screening and talent
will be in the HTML and CSS and javascript." management processes. Overall, the literature highlights
ongoing efforts to develop innovative techniques that
address the evolving needs of recruiters and HR
professionals in the recruitment landscape.
A review of existing tools for resume analysis and job
matching reveals several limitations, particularly with
traditional keyword-matching systems, which can be
manipulated and fail to account for the context of
candidates' skills and experiences. Recent advancements in
NLP and ML offer opportunities to address these issues, but
current systems lack comprehensive solutions that include
both resume analysis and personalized recommendations.

The paper [1]. This study presents a basic text-

processing tool for resume analysis that uses lexical analysis
techniques to extract information from resumes. While the
system can handle simple resume formats, it struggles with
diverse and unstructured formats, making it unreliable for
real-world applications. The research highlights the
challenges of processing unstructured data and concludes
that there is a need for more robust systems capable of
Fig. 2.1. Proposed System handling Complex resume formats. Research Gap: The
system cannot process resumes that vary in structure or
Fig. 2.1.Proposed System The "JobHelper" system is an content, often relying on predefined templates. This creates
AI-driven tool designed to assist users in creating, a significant gap in its ability to parse and analyze resumes
analyzing, and improving resumes, while also offering job effectively across industries.
recommendations and portfolio code generation. Users can
either generate a new resume by selecting a template and The paper [2], This paper introduces a more
entering their information or upload an existing resume for sophisticated resume screening tool that uses advanced NLP
analysis. Uploaded resumes undergo preprocessing to techniques to extract data from resumes with a higher
handle missing data, standardize formats, clean information, accuracy rate (85% for text parsing and 92% for job
and extract key job-related requirements. The system utilizes ranking). The system also provides real-time processing,
AI to refine resumes by generating professional text, allowing recruiters to review resumes faster than traditional
integrating responses into templates, and ensuring the output methods. However, the system's high time complexity and
is polished and ready for job applications. The final resume reliance on older machine learning models, such as Naïve
can be exported in formats such as PDF or Word and is Bayes, make it less efficient in real-world, high-volume
stored in the database for future use. Additionally, the recruitment scenarios. Research Gap: Despite its
system recommends relevant jobs based on the user’s profile improvements over basic text processing methods, this
and may assist in generating portfolio code, streamlining the system struggles with high time complexity and is limited in
entire career preparation process. its ability to rank resumes accurately for roles requiring
nuanced skill matching. Additionally, it does not address the
issue of resume optimization or job recommendations.
The paper [3], This research introduces the concept of but also suggests skills to improve the user's resume score.
integrating resume analysis with a recommender system to Research Gap: Current job recommendation systems often
match candidates with job roles. The paper describes a overlook the detailed skill set of candidates and rely heavily
Python-based application that applies basic keyword on domain interest. This paper addresses this gap by
matching to assess candidate-job fit. While it provides a incorporating a more comprehensive analysis of candidates'
user-friendly interface and has achieved a degree of success skills and projects.
in recommending jobs, it heavily relies on keyword
matching, which can be easily manipulated by including IV. RESULTS
specific keywords in resumes, leading to poor candidate-job-
matches. Research Gap: The system fails to consider the
deeper contextual understanding of candidates'
qualifications, as it relies on simple keyword matching. This
may result in incorrect or irrelevant job recommendations,
especially in cases where the resume contains industry
jargon or ambiguous job titles The literature survey
highlights various strategies for Parkinson's disease (PD)
detection. Wang et al. (2020) excel in early detection,
achieving 96.45% accuracy with a deep-learning method for
premotor PD stages. Prashanth et al. (2018) use patient
questionnaires, achieving over 95% accuracy in
distinguishing early PD from healthy individuals. Fig. 4.1. Initial Interface: Choose Something and Upload Resume

The paper [4], The paper titled "Resume Analyzer Using The image showcases a user interface for a resume analysis
Text Processing" presents a system designed to streamline and job recommendation web application. A user can
the recruitment process by analyzing and ranking resumes upload their resume in various formats, as demonstrated by
using text mining and machine learning techniques. The the "Analyze Test Resume.pdf" file. The system then
system extracts relevant information from resumes, matches analyzes the resume, identifying skills and potential errors,
candidates with job requirements, and ranks them based on to provide tailored job recommendations. The "Choose
the alignment of their skills with the job descriptions, Something..." dropdown hints at user profile management.
thereby assisting recruiters in identifying the best candidates
more efficiently. Research Gap: The primary research gap
identified in the paper is the absence of robust systems
capable of accurately analyzing unstructured resume data
across various formats without heavy reliance on predefined
templates or extensive manual labeling. Existing methods
often struggle with variability in resume formats and require
significant human effort for data annotation, limiting their
scalability and effectiveness. The proposed system aims to
address these limitations by using a combination of text
mining, machine learning, and rule-based approaches to
create a more adaptable and automated resume analysis Fig. 5.2. User Information Form: Enter Personal Details
process.
The image shows the user interface for "JobHelper," an AI-
The paper [5], The paper presents a machine learning powered resume analyzer and job recommender. Users can
and NLP-based system for automating resume screening. upload resumes, as seen with the "Analyze Test
The system aims to address the challenges faced by large Resume.pdf" example, and the system promises smart
companies when screening massive volumes of resumes. recommendations. A "Choose Something" menu and user
Using models like KNN and SVM, helps identify resumes information fields suggest profile management features.
that best match job descriptions and provides suggestions for The interface highlights easy resume upload and AI-driven
improving them. Research Gap: The paper highlights a gap job suggestions.
in existing resume screening systems, specifically that they
do not provide feedback to candidates on how to improve
their resumes. Additionally, most existing systems are
trained once and do not self-improve over time, potentially
limiting their effectiveness for evolving datasets and job
requirements.

The paper [6] The paper presents a job recommendation

system that uses machine learning algorithms to suggest jobs
for college graduates based on their skill sets. The system
analyzes resumes and compares them to job descriptions,
taking into account degrees, soft skills, hard skills, and
projects completed by the user. It not only recommends jobs
 Fig. 4.3. Skills Recommendation: View and Improve Skills Fig. 4.6. Resume Tips: Tips for Enhancing Your Resume

Fig.4.3 This image showcases the "Skills Recommendation"

feature of JobHelper. It displays "Your Current Skills"
extracted from the uploaded resume (likely "Analyze Test
Resume.pdf" as seen before) and "Recommended Skills"
suggested by the AI. The highlighted message, "Our
analysis says you are looking for Web Development jobs,"
demonstrates the system's ability to understand user intent.
The "Courses & Certificates Recommendations" section
further aligns with the project's goal of comprehensive job
search assistance, going beyond just resume analysis and
recommendations. This image represents a key component
of JobHelper's value proposition: providing actionable skill-
based guidance to improve a user's job prospects. Fig. 4.6. Resume Score: Overall Assessment of Resume Quality ATS score

Fig. 4.4. Resume Analysis Results: See Resume Strengths and Weaknesses

Fig 4.4. "Resume Analysis" section of JobHelper. It

displays a parsed resume with basic information extracted,
including name, contact details, degree, and resume pages.
The system has identified the user as being at an
"intermediate level" and is about to present "Skills
Recommendations" based on "Your Current Skills" already
listed. This highlights JobHelper's ability to analyze a
resume, extract key data, assess the user's level, and prepare
to provide tailored skill recommendations for job seekers.

Fig. 4.5. Course Recommendations: Suggested Courses for Skill

Improvemen
 One of the standout features of the system is its use of
Pyresparser and SpaCy for resume parsing and skill
extraction. These tools enable the system to handle a wide
variety of resume formats, including unstructured and
complex ones, which has been a significant limitation in
traditional resume analysis tools. The integration of
Streamlit for the user interface further enhances the user
experience by providing a seamless and intuitive platform
for resume creation and job matching.

VI. CONCLUSION

The project provides a comprehensive solution for job

seekers to analyze their resumes and give them suggestions
on what skills should they add and also suggests some
courses along with the resume ATS score. The system uses a
combination of NLP and machine learning techniques to
provide accurate results. The system has the potential to
improve job search outcomes and provide a competitive
edge to job seekers. The AI-driven resume Analyzer with
Job Recommender and Portfolio Code Generator is a
comprehensive solution designed to streamline the job
application process for modern job seekers. By integrating
advanced technologies such as NLP, machine learning, and
generative AI, the platform offers a range of features that
address the key challenges faced by job seekers, from
resume optimization to job matching and portfolio creation

VII. FUTURE SCOPE

T Analyzer with Job Recommender and Portfolio Code

Generator holds significant potential for future
advancements. Integrating advanced Natural Language
Processing (NLP) techniques can enhance the system's
ability to accurately extract and interpret nuanced
V. DISCUSSION information from user resumes, leading to more precise skill
identification and job matching. Implementing machine
The project aims to provide a comprehensive solution for learning algorithms that learn from user interactions can
job seekers to analyze their resumes, suggest job profiles, provide increasingly personalized job suggestions over time,
and generate portfolio websites. The system uses a improving user satisfaction and engagement. Incorporating
combination of NLP and machine learning techniques to real-time labor market data will ensure that job
analyze resumes and suggest job profiles. The portfolio code recommendations reflect current industry demands and
generator uses a template engine to create customizable emerging roles, keeping users informed of the latest
portfolio websites. The system was tested with a dataset of opportunities. Expanding the portfolio code generator to
100 resumes and achieved an accuracy of 90% in suggesting support additional frameworks and libraries will enable
job profiles. The portfolio code generator was able to create users to create more dynamic and interactive online
functional portfolio websites within minutes. The results portfolios, showcasing their work more effectively.
show that the system is effective in analyzing resumes and Developing a feedback mechanism where users can rate the
suggesting job profiles. relevance of job recommendations and the quality of
generated resumes will allow the system to refine its
The AI-driven resume Analyzer with Job Recommender algorithms based on user input, enhancing its effectiveness.
and Portfolio Code Generator represents a significant Facilitating direct applications by integrating with major job
advancement in the job application process, leveraging portals can streamline the process for users to apply to
cutting-edge technologies such as Natural Language recommended positions, increasing their chances of success.
Processing (NLP), machine learning, and generative AI. The Extending the system's capabilities to analyze resumes and
system's ability to parse resumes, extract key skills, and generate content in multiple languages will cater to a global
match them with relevant job opportunities addresses a user base, making the tool more accessible. Implementing
critical pain point for job seekers, who often struggle to advanced encryption and data anonymization techniques
tailor their resumes to specific job descriptions. By will ensure user data remains secure and private, building
automating this process, the platform not only saves time but trust among users. Establishing a continuous learning
also increases the likelihood of job seekers securing framework where the system adapts to changes in job
interviews and job offers. market trends, user preferences, and technological
advancements will keep the platform relevant and effective.
By pursuing these developments, the system can evolve into [9] ZINJAD, S. B., BHATTACHARJEE, A., BHILEGAONKAR, A.,
LIU, H., “ResumeFlow: An LLM-facilitated Pipeline for Personalized
a more robust, user-centric platform that effectively bridges Resume Generation and Refinement,” arXiv preprint
the gap between job seekers and potential employers. arXiv:2402.06221, 2024
[10] RAHMAN, M., FIGLIOLINI, S., KIM, J., CEDENO, E., KLEIER,
REFERENCES C., SHAH, C., CHADHA, A., “Artificial Intelligence in Career
Counseling: A Test Case with ResumAI,” arXiv preprint
arXiv:2308.14301, 2023.
[1] S. Sahithi, V. Sahithi, T. Sai Akhil, U. Sahithi, G. Abhinay Datta, Prof [11] KULKARNI, A., KENGE, O., MEHATRE, R., KULKARNI, R.,
Maikandhan “A Smart Resume Analyser for Career Optimization “Advancing Job Search: A Comprehensive Resume-Based Job
Using NLP” Recommendation System Using NLP and Deep Learning
[2] NLP - Automated resume analysis and skill suggesting website - Techniques,” International Journal for Research in Applied Science
NUNNA PRASANNA SAI VENKATESH and Engineering Technology, 2024
[3] Ai Resume Analyzer Using Natural Language Processing and Data [12] ZHENG, Z., QIU, Z., HU, X., WU, L., ZHU, H., XIONG, H.,
MiningJayshri Mankar, Abhishek Chauhan, Aniket Gophane, Aditya “Generative Job Recommendations with Large Language Model,”
Karle, Taslimarif arXiv preprint arXiv:2307.02157, 2023
[4] Resume Analyzer Using NLP by Saideepthi et al. (2021): [13] SINGH, A., PAL, S., “Resume Recommendation System Using AI,”
[5] Pratik G. Raut, Prof. Rajesh D. Wagh “Resume Analyzer and International Research Journal of Modernization in Engineering
Recommender System Using Python “ (2021) Technology and Science, vol. 4, no. 7, pp. 1234-1240, 2022
[6] Resume Analyzer Using Text Processing by B.Kelkar1, R.Shedbale2 , [14] DU, Y., LUO, D., YAN, R., LIU, H., SONG, Y., ZHU, H., ZHANG,
D.Khade3, P.Pol4, A.Damame .(2022): J., “Enhancing Job Recommendation through LLM-based Generative
Adversarial Networks,” arXiv preprint arXiv:2307.10747, 2023
[7] Resume Screening Using Machine Learning and NLP: A Proposed
System by Bhushan Kinge, Shrinivas Mandhare, Pranali Chavan
(2022)
[8] JOB RECOMMENDATION SYSTEM BASED ON SKILL SETS by
G.Mahalakshmi1, A.Arun Kumar, B.Senthilnayaki1 , J.Duraimurugan
(2022)