At Sakhi Robotics, I architected a real-time vision pipeline for a hospitality robot, running seven CV/ML models within a 15 ms latency budget. I also engineered an async FastAPI backend with LLM failover so a provider outage no longer broke the assistant.
On my Database and Information Systems course project, I shipped a paper-trading platform on live NSE data, with a React client, FastAPI backend, and PostgreSQL schema. I serialized order fills with row-level locks and exposed REST endpoints alongside a WebSocket stream for live updates.
My systems projects include extending the xv6 teaching kernel and writing a Unix shell in C. I also built a C++17 road-network query engine and a voice-controlled local AI agent with safeguards and tests.
