At Google, I work on productionizing scalable AI/ML and agentic systems for complex enterprise workflows across Google’s People/HR ecosystem. I design multi-agent architectures and LLM reasoning workflows, including planning, tool use, memory, evaluation, and human-in-the-loop orchestration.
At ServiceNow, I built a self-correcting ReAct agent over an enterprise knowledge graph, with a planner-synthesizer architecture and guardrails for grounded answers. I also built federated agentic retrieval across a knowledge graph and live APIs, alongside an LLM-as-judge evaluation framework.
At Hub Platform Technology Partners, I built AI tools that turned natural-language prompts into running apps and Snowflake-style expressions; the expression engine achieved 40% higher accuracy than few-shot prompting and made report creation 30% faster. Earlier at PeopleStrong, I developed machine-learning systems for SQL analytics, chatbots, job matching, attrition prediction, recommendations, and document parsing.

