I've built production-grade LLM systems, RAG architectures, agentic workflows, and scalable inference pipelines for Infosys, TFORM, Meta, and GoHealth over 10+ years.
At Infosys, I build enterprise RAG systems with GPT-4/4o, Azure OpenAI, LangChain, and LlamaIndex, improving retrieval precision through chunking, embedding selection, and re-ranking. I also engineer tool-calling agents, evaluation harnesses, and cost-optimized inference pipelines using Ray Serve, Triton, and ONNX Runtime.
At TFORM, I developed autonomous AI agents and LLM-powered enterprise search with FAISS and Pinecone, while modernizing MLOps with MLflow, W&B, and Azure DevOps. I mentored junior engineers, led technical design reviews, and delivered models for cybersecurity threat detection, forecasting, and anomaly detection.
Earlier, I deployed NLP, multimodal, fraud-detection, and forecasting systems at Meta and built NLP classification, enterprise search, ETL, visualization, and Python microservices at GoHealth. I focus on reliable retrieval, structured outputs, evaluation-driven development, secure deployment, and responsible AI guardrails.
