Hi, I'm Pragya —
building ML systems
& full-stack apps.
B.Tech Computer Science and Engineering at Quantum University, Roorkee (2023–2027). Experienced in end-to-end data analysis and machine learning — from data cleaning and feature engineering to model building, evaluation, and production deployment — alongside full-stack development with React.js and Node.js.
A little about me
I'm a final-year Computer Science student, and if there's one thing that's true across everything I do, it's that I don't like leaving things half-finished. I'm not the person who's content training a model and calling it done, or designing something that never gets built — I want to see it actually work, end to end, in front of a real person. That's shaped how I've spent the last couple of years: picking up machine learning and full-stack development side by side, rather than choosing one over the other.
I learn best by doing rather than reading — which is why most of my time outside class has gone into hands-on work: hackathons, job simulations, internships, and a fair number of things I built just to understand how they work under the hood. I'm fairly disciplined about following through once I start something, and I'd rather spend longer getting a detail right than ship something half-thought-out.
I'm currently in my final year at Quantum University, Roorkee, and actively looking for an SDE or Data Analyst role where I can keep building things that are actually used, not just demoed.
- Degree
- B.Tech CSE
- Institute
- Quantum University, Roorkee
- Batch
- 2023 – 2027
- Looking for
- SDE & Data Analyst roles
What I work with
Languages & ML
- Python
- Pandas
- NumPy
- Scikit-learn
- Matplotlib
- JavaScript
- SQL
ML Techniques
- Regression
- Classification
- Clustering
- Ensemble Methods
- Hyperparameter Tuning
- PCA
Web & Tools
- React.js
- Node.js
- REST APIs
- MongoDB
- Power BI
- Git / GitHub
CS Fundamentals
- DSA
- OOP
- DBMS
- OS
- Computer Networks
Things I've built
CodeBurnout AI
83.7% accuracy- Built an end-to-end burnout-risk scoring engine analyzing developer commit-history data pulled live via the GitHub REST API.
- Engineered 20+ behavioral features from commit timing, frequency, and message sentiment; designed a weighted rule-based scoring algorithm classifying risk into Healthy, Warning, and High Risk tiers.
- Deployed an 11-page modular data application on Streamlit Cloud with automated ReportLab PDF report generation and Plotly visualizations.
Apollo Clinic AI Patient Assistant
- Built on Tars with Google Gemini 2.5 Flash — answers patient FAQs from a clinic knowledge base and identifies New vs Existing patients.
- Captures appointment bookings through a conversational flow and creates them as Leads in Salesforce.
- A second AI Agent scores each completed conversation as a Hot, Warm, or Cold lead and logs the full chat as a Salesforce Task.
AI Resume Screener
- Built a tool that compares uploaded resumes (PDF/DOCX) against a job description and ranks candidates by match score.
- Uses the Groq API (gpt-oss-120b) to extract structured candidate info and generate a 0–100% match score with a detailed breakdown.
- Built a Streamlit interface for pasting a JD, uploading resumes, and viewing ranked results.
Smart Offer Slot Booking System
- Full-stack app where businesses create limited-time offers with bookable time slots and customers book them in real time.
- Built an admin dashboard (offer/slot management, booking status updates, analytics) and a customer-facing booking flow.
- Enforced business rules server-side — slot capacity, expiry, and per-customer booking limits — using Supabase with Row Level Security.
Car Price Prediction
84% R²- Built a regression model predicting resale car prices from historical listing data.
- Performed data cleaning and feature engineering on categorical and numerical listing attributes to improve model fit.
- Compared multiple regression approaches to select the model with the best generalization performance.
Student Depression Prediction System
83.7% accuracy- Built a classification pipeline to predict depression risk from academic and lifestyle survey data.
- Cleaned and preprocessed raw survey data, engineered predictive features, and benchmarked multiple algorithms.
- Evaluated models using accuracy, precision/recall, and confusion-matrix analysis.
Messy Hands Cook-Along
- Redesigned a tablet recipe screen into a zero-touch, hands-free "Cook-Along Mode" for cooks with messy hands.
- Designed a voice- and gesture-driven interface with persistent multi-timer tracking and live ingredient scaling.
- Built a working interactive prototype in HTML/CSS/JS with real browser voice recognition and live countdown timers.
Binance Futures Trading Bot
- Engineered a CLI-based trading bot integrating the Binance REST API, handling authenticated requests, market/limit order execution, and input validation.
- Implemented structured logging and error handling across a modular script architecture.
Weather Application
- Developed a weather application integrating the OpenWeather API to fetch and display real-time weather data.
- Built a responsive UI with client-side form validation and dynamic rendering of location-based results.
Bank Management System
- Command-line bank management system in Python with account creation, deposit, withdrawal, update, and delete.
- Used JSON for persistent data storage.
Where I've worked
Web Design & Leadership Facilitator — Nobel Learning PBC (Virtual)
- Completed a 90-hour Learner-to-Leader training program covering web design, internet troubleshooting, and technical communication.
- Facilitated peer learning sessions, building leadership, teamwork, and collaborative problem-solving skills.
Full Stack Web Development Intern — Anonic Technologies Pvt. Ltd. (Virtual)
- Built and optimized full-stack features using React.js and Node.js, contributing to frontend components and backend services.
- Designed REST APIs and contributed to database schema design.
- Collaborated with a distributed engineering team to debug issues and ship features against deadlines.



















