Subasri Sridhar

I enjoy working where research and development meet production, especially turning ideas into product capabilities.

My work is backend-focused, spanning AI feature integration, tool and MCP-based workflows, and the engineering needed to connect these systems into products.

Areas of expertise include LLM inference and serving, agent engineering, and retrieval systems.

My interests include building and fine-tuning AI systems, reliable and self-recovering agents, deep reinforcement learning, sequential decision-making, and transferable and sample-efficient learning.