At Ford Motor Company, I engineered a production LangChain and GPT-4 pipeline that translated legacy workflows into BigQuery SQL and Dataform DAGs, cutting ETL migration effort by about 80%. I also built an evaluation harness to catch SQL-translation and schema-match regressions before production cutover.
I built a LangGraph debug agent with Claude and GPT-4 that cut issue triage from hours to minutes and saved 1–2 days per monthly close. In my open-source projects, I’ve worked on a governed text-to-SQL agent and an FCRA-compliant credit-risk explainer with checks against certified model reasons.
