End-to-end data pipeline for 1,000+ federal research grants. Star schema, 25+ tests, daily Airflow DAG.

Ingests 1,000+ federal research grants from NSF and NIH APIs into a DuckDB warehouse with dual persistence (database + JSON archive), rate limiting, and per-record quality scoring. Modeled the warehouse as a star schema with 14 dbt models (2 staging, 5 dimensions, 1 fact, 5 marts) and 25+ automated tests covering uniqueness, referential integrity, and value ranges. Orchestrated via Apache Airflow on Docker Compose (4 services) with an 8-task daily DAG, retry logic, and DuckDB single-writer concurrency management.

Quizlet-style flashcards, dockerized, Agile'd, tested.

Collaborated with a team of 4 to build a Quizlet-style flashcard web app. Containerized with Docker, CI via GitHub Actions, tests in Jest, APIs tested with Postman. Agile delivery.

Work in progress. Details are under wraps for now.

████████ ██████ ████ ██████████ ███ ████ ████████. ██████ ████ ███████ ██████. [redacted until launch]

Image recognition for pantry management. 80% accuracy on realtime item detection.

Team of 5. Image-detection pantry app that notifies users of near-expiry items and suggests recipes. Used Roboflow + YOLOv8. Flask backend with SQLite. Built a REST API for data manipulation.

Won top database web project (>90% score). Heroku + Docker auto-deploy.

E-commerce app with shopping + admin flows. Deployed on Heroku via Docker. Auto-deploy on main branch push. Won the Database Web Development Award.