At NetApp, I define the RAG architecture and guardrails for NetAIChat, connecting enterprise documents to Azure OpenAI through controlled retrieval, source grounding, and content filtering. I also design Text-to-SQL safeguards and evaluate systems for hallucination, factuality, bias, and data exposure.
At AWS Annapurna Labs, I developed C/C++ Linux systems software and Python validation workflows for ML accelerator servers, including Trainium2. Earlier at AMD, I contributed to yield improvement across UltraScale+ products and automated silicon analysis with Python and Snowflake SQL.
