At Goldman Sachs, I built Python ETL pipelines in Apache Airflow to refresh financial document data weekly, with alerts to catch failed loads. I also used Apache Spark and Pandas to turn tens of thousands of unstructured documents into datasets.
I helped migrate services and data flows from on-premise servers to AWS, keeping internal data tools running for research analysts. I built and maintained PostgreSQL models and queries behind tools served through FastAPI and GraphQL APIs and React dashboards.
Working with research analysts, I prepared datasets and built classification and similarity models, improving accuracy by about 20%, and retrieval pipelines that raised search relevance by 28%. I monitored services and pipelines with CloudWatch, Prometheus, and Grafana, and automated deployments with Docker, Kubernetes, and GitHub Actions CI/CD.
On TrialBridge, I built an ingestion pipeline that parsed 103,509 ClinicalTrials.gov XML files into a PostgreSQL model of 2,344 studies, with pgvector indexes returning the top 50 matches in about 1 ms. I also added a Redis task queue and an eligibility rules engine.
