Elias Benbourenane

I develop comprehensive tools for biomedical research using React, SAS, and FastAPI, including a platform to explore drug-disease relationships across 13M+ PaTH Network patient records. Recently built an LLM-powered Next.js application with Prisma to visualize and summarize biomedical knowledge graphs of drug mechanisms, and prototyped an ML tool for UPMC hospital to identify dangerous patient care transitions using FastAPI, SQLAlchemy, and React.

A major focus has been modernizing legacy systems and optimizing performance. I modernized a SQL analytics codebase with modular design and incremental processing, reducing runtime by 9x (38h → 4h). Additionally, I've orchestrated parallelized ETL pipelines using SLURM and Amazon Redshift that reduced data processing time by over 90%.

On the infrastructure side, I architected and deployed a secure, HIPAA-compliant Linux and Docker environment for AI compute workloads that now serves 200+ users. I've managed, architected, and deployed 10+ production servers on-premises or in the cloud using AWS EC2/ECS/ALB, often leading architecture decisions and serving as the lead or sole developer across 9+ concurrent projects.

I also contributed with CMU to ship a PyTorch ML framework for predictive temporal modeling through a Svelte web interface, and authored a monitoring framework in Bash with Slack alerting for several server groups critical to research infrastructure.

Promoted post-internship to a consulting role, advising on technical direction and key engineering decisions for multiple projects. I led a comprehensive full-stack migration to a Django and React architecture, introducing new features while enabling faster development and significantly improving code maintainability.

I developed an analytics dashboard using Django and Chart.js to analyze user engagement patterns and inform data-driven outreach decisions. Beyond development work, I implemented robust security measures including a user authentication and authorization system with OAuth 2.0 integration, and established CI/CD pipelines using GitHub Actions that streamlined React builds, Jest tests, and deployment processes on AWS.

I worked on high-performance systems for physics research, focusing on database optimization and custom tooling development. My primary achievement was refactoring SQL tables containing billions of rows of sensor data, which reduced the overall database size by over 40% while significantly improving indexing performance for faster queries.

I extended a custom Grafana fork with new Go and React modules specifically designed for querying high-frequency physics time series data. To support the team's development workflow, I developed comprehensive CI/CD pipelines using AWS, Docker, and GitHub Actions that fully automated daily codebase builds and deployments. Additionally, I engineered automated Go-based SQL data validation pipelines that eliminated approximately 6 hours of manual physicist review work each week.

I operated the university's help desk phone line, providing comprehensive technical support for software, hardware, and general IT issues to students, faculty, and staff. Averaging over 15 calls per day, I assisted users with varying levels of digital literacy, successfully resolving complex technical problems while maintaining clear, patient communication. My focus on user experience and problem-solving resulted in a 100% satisfaction rating throughout my tenure.