CS Graduate · Python · ML · Full Stack · Compilers & Systems
📍 Indore, India | 🎓 University of Arizona | 🔍 Open to Work
- 🧠 Built ML pipelines for computer vision — CNNs, RNNs, OpenCV, PyTorch
- 📈 Building a quantitative finance toolkit for Indian equity portfolio analysis, feature engineering, and benchmark construction
- 🔧 Wrote a full compiler for a Java-like language (Tau) from scratch in Python
- 🌐 Full stack experience: Node.js/Express backends, React frontends, MongoDB, deployed on Linux
- 🐳 Comfortable with Docker, Kubernetes, and Linux Cron automation
- ♟️ Built a multiplayer chess app with Stockfish AI integration
Languages
ML & Data
Web & Backend
Tools & Infra
Undergraduate Research Development Fellow · University of Arizona (Feb 2021 – May 2022)
- Built a Python ETL pipeline using the Asana API, eliminating 10+ hours of manual data entry weekly
- Tracked 500+ research funding opportunities across large grant databases
Undergraduate Research Assistant · Dept. of Astronomy, University of Arizona (Aug 2020 – Jan 2021)
- Migrated a legacy C/Python 2/AWK/Bash astronomy codebase to Python 3, improving performance by 18%
- Automated data processing & graph generation every 4 hours via Linux Cron jobs
- Containerized scripts with Docker for cross-platform compatibility
📈 Stocks · Python · Pandas · yfinance · NumPy · 🚧 Under Construction
Production-structured Python package for Indian equity portfolio analysis. Covers the full pipeline from NSE/yfinance data ingestion and cleaning, through feature engineering (daily returns, annualised return & volatility, MAD-based winsorization) and universe management (symbol removal, weight-threshold filtering). Includes a dedicated scripts layer that builds cleaned, deduplicated historical total-return series for Nifty 500, Nifty Next 50, and Nifty Smallcap 250 as benchmark proxies. Portfolio optimisation (mean-variance, efficient frontier, risk analytics) is on the roadmap.
🪖 Helmet Detection · CNN · PyTorch · OpenCV
CNN model trained on ~600 images achieving 84.4% accuracy and ROC-AUC of 0.812.
RNN classifying COVID-19 from chest X-rays with 79.5% accuracy.
Multiplayer chess with online play, game save/reload, and Stockfish AI integration.
End-to-end compiler for a Java-like language — semantic analysis, type checking, and code generation on top of an LL(1) parser.
Real-time microblogging platform deployed on a DigitalOcean Linux droplet.
MVC-based gym management app with scheduling, billing, and conflict prevention.