At Bharati Vidyapeeth College of Engineering, I architected production AI service backends using FastAPI, Redis, and Docker, with an 89.2% accuracy benchmark. I reduced model inference latency by 70% through 4-bit AWQ quantization and KV-cache optimization.
As a contributor to Instructor (Structured Outputs for LLMs), I authored PyTest test suites and integrated MyPy static type validation. I also refactored API response deserialization routines to reduce edge-case runtime failures across upstream providers.
At Petals NGO, I engineered automated ETL pipelines with Python, PostgreSQL, and AWS S3, cutting reporting time from two days to four hours. I also deployed a real-time inventory allocation dashboard that reduced supply stock-outs by 40%.
My projects include Market Mastermind, a market data simulation engine with order execution and risk profiling, and Arogya Saathi, a distributed microservices healthcare platform. I also developed Ignar API Shield, a FastAPI gateway that detects shadow and unmapped APIs across banking endpoints.
