Seth Ferreira Donaldson’s Post

We're creating an AI-powered resume screener app called "SkillScan" using Python, NLP, and ML. Our project aims to revolutionize the hiring process by developing a sophisticated resume screening application that leverages the power of Artificial Intelligence and Natural Language Processing. By automating the initial stages of talent acquisition, we seek to significantly reduce time-to-hire and improve the overall efficiency of recruitment for HR professionals. Our app will utilize advanced machine learning algorithms to accurately match candidate skills and experience with job requirements, enabling recruiters to identify top-tier talent more effectively. Through intelligent resume parsing and analysis, the application will extract essential information and provide actionable insights to streamline the evaluation process. By combining cutting-edge technology with a user-friendly interface, we are committed to creating a tool that empowers recruiters to make data-driven decisions and focus on building high-performing teams. Checkout the project -> https://lnkd.in/gzxVBuTJ

#Technology #DataAnalytics #DataDriven AI, Write and Style My CV: Created with Midjourney by Author Agentic Workflow Writing Structured Documents, A Line-by-Line Tutorial Not only the Mona Lisa and the Vitruvian Man but also the Curriculum Vitae (CV), are cultural artifacts by Leonardo Da Vinci’s hand that resonate and reproduce in the present time. The CV is not the exclusive way to present oneself to the job market. Yet the CV persists despite the many innovations in information and graphics technology since Leonardo enumerated on paper his skills and abilities to the Duke of Milan. In high-level terms, the creation of a CV: * summarizes past accomplishments and experiences of a person in a document form, * in a manner relevant to a specific audience, who in a short time assesses the person’s relative and absolute utility to some end, * where the style and layout of the document form are chosen to be conducive to a favourable assessment by said audience. These are semantic operations in service of an objective under vaguely stated constraints. Large language models (LLMs) are the premier means to execute semantic operations with computers, especially if the operations are ambiguous in the way human communication often is. The most common way to date to interact with LLMs is a chat app — ChatGPT, Claude, Le Chat etc. We, the human users of said chat apps, define somewhat loosely the semantic operations by way of our chat messages. Certain applications, however, are better served by a different interface and a different way to create semantic operations. Chat is not the be-all and end-all of LLMs.https://https://lnkd.in/g5HpzxSR I will use the APIs for the LLM models of Anthropic (especially Sonnet and Haiku) to create a basic application for CV assembly. It relies on a workflow of agents working in concert (an agentic workflow), each agent performing some semantic operation in the chain of actions it takes to go from a blob of personal data and history to a structured CV document worthy of its august progenitor… This is a tutorial on building a small yet complete LLM-powered non-chat application. In what follows I describe both the code, my reasons for a particular design, and where in the bigger picture each piece of Python code fits. The CV creation app is a useful illustration of AIs working on the general task of structured style-content generation. Before Code & How — Show What & Wow Imagine a collection of personal data and lengthy career descriptions, mostly text, organized into a few files where information is scattered. In that collection is the raw material of a CV. Only it would take effort to separate the relevant from the irrelevant, distill and refine it, and give it a good and pleasant form. Next imagine running a… #MachineLearning #ArtificialIntelligence #DataScience

🚀 Choosing the Right LLM for Resume and Job Application Processing I recently faced the challenge of selecting a suitable large language model (LLM) for my side project. It's an application where I tested AI for HR tasks, requiring processing and matching resumes and job vacancies via API, with the results displayed in a table on a client-side frontend. The decision came down to balancing quality, token limits, and cost. Initially, I tried GPT-4, GPT-4-turbo, and GPT-3.5-turbo. While the quality and pricing of GPT-3.5-turbo were reasonable, it struggled with large amounts of resumes due to token limit restrictions (4,096 tokens per request). Other OpenAI models were too expensive for this task. That led me to explore other options, specifically testing models from Anthropic: Claude 3 Opus, Claude 3.5 Sonnet, and Claude 3 Haiku. Their key advantage? 200,000 tokens per request, which turned out to be a game-changer for handling large volumes of text. However, not all models are created equal. While the Claude Haiku model offered an attractive price point, its analysis results were somewhat mediocre and often produced "hallucinations." On the other hand, Claude 3.5 Sonnet struck the perfect balance between quality and cost, making it the optimal choice for my needs. In the end, the cost of processing 1 job vacancy and 7 resumes was: Claude 3 Haiku: <$0.01 Claude 3.5 Sonnet: $0.04 I'm eagerly waiting for Anthropic to release its new 3.5 Haiku model. I believe this could significantly reduce the costs for simple API requests. My conclusions: 1️⃣ Finding the right LLM isn’t just about picking the most advanced model—it’s about understanding the specific requirements of your project and finding the best fit. 2️⃣ If you refine your prompt and test the results deeply, you can step down to a weaker and cheaper model, which will help save a lot on request costs. 3️⃣ If the output quality of cheaper models is unsatisfactory, you can use them for testing and switch to more powerful ones for production. #LLM #Frontend #AI #NaturalLanguageProcessing #ClaudeAI #GPT4

🚀 Excited to Share My Latest Project! 🚀 Just completed building a Resume Classification & Matching System using Python and Flask! 🎉 This web app extracts text from PDF, DOCX, and TXT files and uses Natural Language Processing (NLP) and Machine Learning to classify resumes based on skills and experience. Some Key Features: - Resume classification for roles like Data Scientist, Software Developer, NLP Specialist, and more - Utilizes pdfplumber, docx, nltk, and sci-kit-learn for powerful text processing and analysis - Built with a Flask backend and a simple frontend to upload resumes and job descriptions Check out the project if you're looking to simplify the hiring process or automate resume screening! 😊 #Python #Flask #NLP #MachineLearning #ResumeScreening #TextProcessing #CareerDevelopment #HiringTech #DataScience #SoftwareDevelopment

Check out how my ATS system uses Google Gemini to match resumes with job descriptions – all from a simple PDF upload! In this project, I’ve built an AI-powered Applicant Tracking System using Google Gemini. It automatically processes resumes uploaded in PDF format, converts them into text, and intelligently matches candidates with the most relevant job descriptions. 👇 Watch the Demo Here! Key Features: Upload resumes in PDF format and have them automatically converted into text. AI-powered matching that compares resumes with job descriptions and identifies the best-fit candidates. Built using Google Gemini, ensuring fast and accurate text extraction and comparison. Designed to save recruiters time by automating manual screening tasks. The system leverages Google Gemini for natural language processing to extract text from PDF resumes and match them with job descriptions. Using Python and libraries like PyPDF2 for PDF parsing, the matching algorithm was fine-tuned to prioritize relevant skills and qualifications. 🔍 I'd love to hear your thoughts! Feel free to share your feedback in the comments, or reach out if you're interested in testing the system. #AI #GoogleGemini #ApplicantTrackingSystem #JobMatching #RecruitmentTech #NLP #AIinHiring #ProjectDemo #ATS #Automation

Imagine taking 600K+ job postings—primarily US tech jobs—and extracting and visualizing insights from raw, unstructured text in real-time. These are the early fruits of an NLP project I started over two years ago. My goal was to understand what skills I needed to invest in to achieve my career goals. I couldn't depend on other tech skill reports because: (a) the BLS and gov are not exactly synced with industry and (b) the reports worth reviewing were lagging behind and usually came out only once a year. Enter Hazon Insights: a visual job market research tool. Hazon Insights leverages natural language processing techniques to transform raw job postings into actionable intelligence. By continuously analyzing over 600,000 U.S. tech job postings, it uncovers real-time trends in skill demand, salary benchmarks, and emerging technologies across the tech industry. What Makes Hazon Insights Stand Out? Real-Time Data with Historical Context: Access salary ranges and years of experience segmented by company, job titles, skills, credentials, domains, subdomains, and specializations—all updated in real-time and enriched with historical values. Advanced Salary and Skill Analytics: Top Skills and Compensation: Analyze the 90th percentile of job postings for skill count to identify which skills are most sought after and which command the highest salaries. Skill Demand Visualization: View the top tech skills plotted by average salary (x-axis) versus average years of experience (y-axis) in an interactive chart, helping you pinpoint lucrative skills relative to experience levels. Canonicalized Skill Counts: Explore standardized skills with occurrence counts to understand their demand across the industry. Interrelated Skills: Discover how skills are interconnected across various specializations, aiding in comprehensive skill development. Interactive Location Filtering: Utilize city, state, and remote filters on a map to focus on specific geographical areas that matter to you. Specialization Trends: Examine salary trends over years of experience within specific specializations, extracted directly from job postings to spot growth opportunities. Quick Access to Job Details: Instantly view job posting responsibilities, qualifications, and more, giving you immediate insights into what employers are seeking. Explore the trends shaping the future of tech. Visit www.hazon.fyi and unlock the data you need to succeed.

#AI #ML #Tech AI, Write and Style My CV: Created with Midjourney by Author Agentic Workflow Writing Structured Documents, A Line-by-Line Tutorial Not only the Mona Lisa and the Vitruvian Man but also the Curriculum Vitae (CV), are cultural artifacts by Leonardo Da Vinci’s hand that resonate and reproduce in the present time. The CV is not the exclusive way to present oneself to the job market. Yet the CV persists despite the many innovations in information and graphics technology since Leonardo enumerated on paper his skills and abilities to the Duke of Milan. In high-level terms, the creation of a CV: * summarizes past accomplishments and experiences of a person in a document form, * in a manner relevant to a specific audience, who in a short time assesses the person’s relative and absolute utility to some end, * where the style and layout of the document form are chosen to be conducive to a favourable assessment by said audience. These are semantic operations in service of an objective under vaguely stated constraints. Large language models (LLMs) are the premier means to execute semantic operations with computers, especially if the operations are ambiguous in the way human communication often is. The most common way to date to interact with LLMs is a chat app — ChatGPT, Claude, Le Chat etc. We, the human users of said chat apps, define somewhat loosely the semantic operations by way of our chat messages. Certain applications, however, are better served by a different interface and a different way to create semantic operations. Chat is not the be-all and end-all of LLMs.https://https://lnkd.in/g5HpzxSR I will use the APIs for the LLM models of Anthropic (especially Sonnet and Haiku) to create a basic application for CV assembly. It relies on a workflow of agents working in concert (an agentic workflow), each agent performing some semantic operation in the chain of actions it takes to go from a blob of personal data and history to a structured CV document worthy of its august progenitor… This is a tutorial on building a small yet complete LLM-powered non-chat application. In what follows I describe both the code, my reasons for a particular design, and where in the bigger picture each piece of Python code fits. The CV creation app is a useful illustration of AIs working on the general task of structured style-content generation. Before Code & How — Show What & Wow Imagine a collection of personal data and lengthy career descriptions, mostly text, organized into a few files where information is scattered. In that collection is the raw material of a CV. Only it would take effort to separate the relevant from the irrelevant, distill and refine it, and give it a good and pleasant form. Next imagine running a script make_cv and… #MachineLearning #ArtificialIntelligence #DataScience

🚀 Autonomous AI-Powered Resume Processor and Mock Interview System 🎯 I recently developed an AI-powered autonomous system designed to streamline resume processing and conduct mock interviews with real-time sentiment analysis. This project integrates cutting-edge technologies to assist job seekers in preparing for interviews, while also providing detailed feedback to help them improve. Key Features: Automated Resume Parsing: Leveraged pdfplumber and docx2txt for seamless extraction of text from both PDF and Word files. Integrated a Named Entity Recognition (NER) pipeline using a pretrained BERT model to extract key information such as skills, experience, and education from the resume content. Intelligent Question Generation: Dynamically generates customized interview questions based on extracted skills, past experiences, and educational background. Supports role-specific question generation (e.g., Data Scientist, Software Engineer) and allows for the inclusion of custom questions. Voice Interaction and Real-Time Sentiment Analysis: Utilized pyttsx3 for Text-to-Speech (TTS) and speech_recognition for Speech-to-Text (STT), enabling voice-based interviews with users. Implemented Hugging Face's sentiment analysis pipeline to evaluate interview responses in real time, providing feedback on sentiment and confidence. Interview Summary: After the interview, a detailed summary is generated, which includes the questions asked, user responses, and sentiment analysis, helping users track their progress and areas of improvement. This project showcases the potential of AI and NLP in revolutionizing the recruitment process by providing an efficient way to conduct autonomous interviews, evaluate responses, and give candidates personalized feedback. Tech stack highlights: NLP with Hugging Face Transformers (BERT, Sentiment Analysis) PDF and Word file handling with pdfplumber and docx2txt Speech interaction using pyttsx3 and speech_recognition Python async capabilities for smooth, concurrent processing Excited to explore more AI-driven innovations in recruitment and beyond! 💡 #AI #NLP #ResumeParser #MockInterview #Automation #JobSearch #MachineLearning #TechInnovation

🚀ATS Resume Expert - Cutting-Edge Application Using Generative AI 🚀 This project is a state-of-the-art Application Tracking System (ATS) powered by LangChain, Google Gemini LLM , and Streamlit. It offers an intelligent solution for job seekers in tech fields like Software Engineering, Data Science, Fullstack Development, Big Data Engineering, DevOps, Machine Learning (ML), Deep Learning (DL), and Generative AI Engineering, helping them optimize their resumes for the competitive job market. Key Benefits: 🌟 AI-Powered Resume Evaluation: The system uses Google Gemini to automatically evaluate resumes based on a given job description. It provides detailed feedback, including: 📊 Percentage match between the resume and the job description. 🔑 Identification of missing keywords that can increase the chances of getting shortlisted by ATS. 📝 A profile summary that outlines how the candidate's skills align with the job's requirements. Personalized Resume Optimization: Tailors recommendations based on the job description provided by the user, offering advice on missing skills or areas to improve, making it easier to align resumes with job postings. Time Efficiency: Job seekers can quickly upload a resume and receive instant, automated feedback, reducing the need for manual revision and external consultation. ⏱️ Improved ATS Compatibility: With many companies relying on ATS to filter resumes, this tool helps users optimize their resumes to pass ATS filters and improve their chances of landing an interview. ✅ User-Friendly Interface: Built with Streamlit, the app provides a simple and intuitive interface for uploading resumes, inputting job descriptions, and getting results in real time. 🖥️ Problems Solved: 🛠️ ATS Screening Issues: Many resumes get rejected by Applicant Tracking Systems due to missing keywords or improper formatting. This app helps users identify what their resumes are missing, thus improving their chances of passing the initial screening. ❌➡️✅ Job Application Optimization: Job seekers often struggle with tailoring their resumes to different job descriptions. This application automatically highlights areas where the resume can be improved, helping users make their resumes more competitive. 🎯 Resume Writing Guidance: By generating a profile summary and matching score, the app gives users clear direction on how to improve their resumes, saving them from guesswork. 📖 Increased Confidence for Applicants: Knowing how well their resumes align with a job description, users can apply for jobs with more confidence, leading to a better job search experience. 💪✨ Check the application here:https://lnkd.in/dUkuk4sT #GenerativeAI #ArtificialIntelligence #MachineLearning #DataScience #ResumeOptimization #JobSearch #CareerDevelopment #ATS #Streamlit #LangChain #GoogleGemini #NLP #TechCareers #FullstackDevelopment #DevOps #BigData #SoftwareEngineering

Introducing the Resume Category Classifier, a tool designed to automatically classify resumes into predefined categories with high accuracy. Here's a breakdown of the process: 1️⃣ Text Data Preprocessing: We start by cleaning and preprocessing the resume text data using the NLTK library. This step includes removing stopwords, stemming, and other text normalization techniques to ensure the data is clean and ready for analysis. 2️⃣ Vectorization with TF-IDF: The preprocessed text data is then converted into vectors using the Term Frequency-Inverse Document Frequency (TF-IDF) method. This transformation allows us to represent the textual data in a numerical format that can be effectively used for classification. 3️⃣ Supervised ML Training: We train the model using a supervised machine learning algorithm. This involves feeding the labeled data into the algorithm so it can learn to distinguish between different resume categories based on the features extracted in the vectorization step. 4️⃣ User Interface with Streamlit: To make the tool user-friendly, I have implemented an interactive web interface using Streamlit. Users can easily upload resumes, and the system will automatically classify them into categories such as 'Software Development', 'Data Science', 'Marketing', etc. This project showcases the practical application of natural language processing and machine learning, aiming to make the resume sorting process more efficient and effective for recruiters and HR professionals. I am eager to hear any feedback or suggestions you might have. Let's innovate and transform the recruitment process together! 💼🤖 #DataScience #MachineLearning #NLP #Recruitment #AI #Streamlit #Python #CareerDevelopment #Innovation