● Scripted SAS and SQL programs for internal use in the PCORnet database, a national medical network of hundreds of hospitals and over 30,000,000 patients yearly, that generate statistical summaries of patient electronic health records. ● Designed a medical informational tool with React.js and FastAPI for millions of PCORnet patients to better understand their conditions/medications through charts and tables with powerful querying capabilities. ● Desgned a dashboard with React.js for care coordinators at UPMC hospital to predict prescription errors during cross-center patient transitions with machine learning which has been used to retroactively identify 100+ dangerous failure modes in recent transitions. ● Developed a REST API for use in medical research to query various AI language models in lieu of a shared GPU environment that is optimized with block-wise quantization and request queuing to manage high traffic and concurrent requests. ● Deployed a high-performance Linux server for research on large medical datasets packaged with locked-down Jupyter and RStudio environments in user-isolated and GPU-enabled Docker containers. ● Wrote Bash scripts to automatically create Jupyter/RStudio Docker environments, eliminating 120+ minutes of manual build time.

● Migrated the 412Connect project, a service to educate and inform the public about local businesses, to an entirely new stack, handling backend, frontend, and database design in a Django, React.js, and SQLite stack. ● Developed a user analytics and tracking system for 412Connect, featuring a dashboard, charts, graphs, and search/filtering. ● Introduced new DevOps pipelines, including CI/CD building, testing, and deployment to AWS using GitHub Actions, eliminating 90+ minutes of build, test, and deployment time from the daily development workflow of the team. ● Promoted to a consulting role to provide guidance and strategic direction for the project beyond my internship tenure.

● Refactored SQL tables containing billions of rows, reducing the database's size by over 40%, optimizing indexing performance. ● Maintained a Grafana fork, involving full-stack development with Go and React.js to integrate features for TimescaleDB. ● Wrote GitHub Actions pipelines to automate the building and deployment of Docker images for AWS EC2 containers, eliminating 15+ minutes of build and deployment time daily. ● Served as a webmaster, handling DNS management, SSL certificate installation, and web hosting for the lab's websites.

● Operated a help desk phone line, providing support for software, hardware, and non-technical issues, averaging 20+ calls a day. ● Documented all incoming calls, emails, and live chats with detailed solutions and troubleshooting steps taken to solve issues.