At Team Fateh Formula Racing, I develop C-based data acquisition firmware for microcontrollers that processes, filters, and logs high-frequency sensor telemetry. I also implemented serial streaming for live data transmission and real-time system monitoring.
For FormZero, I engineered FastAPI endpoints and a semantic RAG pipeline for real-time eligibility analysis across US government programs. I also developed a backend calculation engine for retroactive missed-benefit values and added PII redaction and per-IP rate limiting.
On my Mini Agentic Code Verifier project, I built a LangChain and Groq API pipeline that generates, executes, and validates Python solutions. Its self-correcting loop uses runtime traceback context to help debug generated code.
I built a Hindi voice-to-voice RAG system and deployed it on Hugging Face Spaces. By switching to a multilingual embedder and using INT8 quantization, I reduced ASR latency by 45%.
