An AI speaking partner for English learners: voice conversations, AI-scored mock interviews, and Hindi-to-English practice. Replies come from AI language models via Groq, Google Gemini and OpenRouter.
Backtesting and paper trading for Indian stocks, indices and crypto, on 8 years of 5-minute NSE equity data (2018–2026, 428 symbols) with a realistic cost model — brokerage, STT, exchange charges, GST and slippage. No signals, just proof.
In development — YAHIN, a local-commerce product that shows whether a shop is open and has what you need before you go.
What each case study covers
These aren’t screenshots with a stack list. Each one explains a real product the way an engineer would review it — so you can judge how I think, not just what I shipped.
The problem, and who it was for.
The architecture, drawn as a layered diagram, with every layer explained.
The decisions and trade-offs — what was chosen, what it cost, and why.
What failed or changed along the way, and what I learned.
Both products were built end to end — product, frontend, backend, the AI layer, data and deployment. If you’re planning something similar, here’s how we build AI products.