AI Engineer · Full-Stack Developer · Cricket Enthusiast
I build things at the intersection of AI systems and real-world applications — from multi-agent pipelines to personal finance tools for Indian investors.
- Indian Portfolio Tracker — A self-hosted, offline-first personal portfolio tracker for Indian investors. Supports 12+ asset classes, CAS PDF import, XIRR, tax harvesting & goal tracking. Stack: FastAPI · React/TypeScript/Vite · SQLite
- CricketIQ — Fantasy Best XI IPL selector built on LangGraph with parallel stat agents (venue, head-to-head, recent form) and a Groq-powered selector node. Live on Streamlit.
- Neural Networks in Manim — Animated educational video series on neural networks (based on Karpathy's micrograd), targeting Python developers with zero ML background.
AI / ML
LangGraph LangChain PyTorch
Backend
Python FastAPI SQLite SQLAlchemy Node.js
Frontend
React TypeScript Vite Streamlit
Tooling
GitHub Actions Docker Manim
- PyTorch — working through Andrej Karpathy's course with hands-on projects
- Agentic AI patterns — multi-agent architectures, A2A protocols, LangGraph orchestration
I'm an RCB fan who ended up building an AI-powered IPL Fantasy selector. Coincidence? Definitely not.
I love combining personal passions with technical depth — cricket data, Marathi cultural community work, and production-grade AI systems all find their way into my projects.
Building in public. One agent at a time.