Aakash Vijeta is a Data Science & AI undergraduate at IIT Guwahati building production machine-learning systems, quant trading research and reinforcement-learning agents.
Machine Learning — A calibrated machine-learning service that calls every Formula 1 podium before the lights go out — with a public accuracy scoreboard to keep it honest.
Built with Python, LightGBM, Optuna, scikit-learn, FastAPI, FastF1, Postgres, SQLite, React 19, Vite 8, GSAP.
Quant Research — A Bloomberg-terminal-styled pre-market trading desk for the Nifty 50 — walk-forward validated XGBoost signals, a Groq-powered AI analyst, and a dashboard that lands on your screen before the opening bell.
Built with Python, XGBoost, pandas, FastAPI, React 19, Vite 6, Framer Motion 12, Netlify Edge, Groq · Llama 3.3 70B, OpenRouter.
Reinforcement Learning — A Soft Actor-Critic agent that teaches itself to drive fast laps around real Formula 1 circuits — and finishes 48 of 50 tracks it has never seen.
Built with Python, Stable-Baselines3, SAC, Gymnasium, NumPy, pygame, ffmpeg, TensorBoard, pytest, GitHub Actions.