At h2olabs.io, I architect RAG and agentic AI systems, including vector search with Elasticsearch and Pinecone, real-time LLM services with FastAPI and WebSockets, and multi-step workflows on Google Cloud Run. I also deliver speech intelligence, privacy-preserving audio, IoT anomaly detection, and scalable computer vision pipelines.
Across CodeDistrict and WebTech, I've productionized ML models and APIs using Docker, AWS, PostgreSQL, Celery, Redis, and RabbitMQ. My work spans YOLO-based detection, sports video intelligence, NLP, time-series forecasting, and deep-learning inference optimization that improved performance from 5 FPS to 33 FPS.

