ai engineer at asurion and data science grad student at USF, based in san francisco.
i build RAG pipelines, search systems, evaluation frameworks, and AI agents for production.
- payroll agent - turns a messy payroll email into human-approved paychecks; deterministic code (not confidence scores) gates every money decision, with real IRS Pub 15-T withholding
- gtm research pipeline mcp server - scores company domains against an editable ICP and drafts outreach where every claim traces to a cited retrieval or gets dropped. Groundedness measured by an LLM-as-judge eval; live as a hosted MCP server
- sf date night concierge - building an agentic RAG pipeline that plan dates in SF and making reservations on your behalf
- customer service bot intent detection - e2e voice pipeline (stt -> tts) and llm orchestration for intent detection
- multi-tenant RAG pipeline (asurion) - ingestion and retrieval system with zero-code model deployment
- KB health detection (asurion) - corrective RAG system that flags outdated knowledge base content
- weather prediction pipeline - daily Airflow pipeline on GCP that predicts tomorrow's high temp
- nba home court advantage - statistical deep dive into whether it still matters
python, sql, langchain, rag, llms, bigquery, airflow, mongodb, gcp, docker, git, anthropic, pydantic, pytest, exa, browserbase