Shamit Billa

I build systems which are fast, reliable and correct—from low-latency trading exchange engineering in C++ to agentic retrieval pipelines at enterprise scale. I’m currently an AI Engineering Intern at Cohesity, where I’ve delivered an enterprise Pricing Intelligence Agent with recursive ingestion and a two-stage retrieval architecture to reduce hallucinations and speed up pricing query resolution.

I also enjoy end-to-end AI infrastructure: I built an AI code documentation system that auto-generates retrievable knowledge chunks from internal codebases and serves it via a Claude MCP server. Beyond production work, I’m researching intent-aware LLM safety using multidimensional signal decomposition, benchmarking safety classifiers, and pushing for defenses that detect “why” behind prompts—not just “what” users ask.