RESUME

Building automation workflows and AI-powered systems for an AI gifting platform — integrating APIs, LLM services, and data pipelines into production infrastructure.

Designing and deploying end-to-end automation that supports product, sales, and marketing functions, reducing manual work across the team.

Own end-to-end delivery of QuitTxt clinical AI platform serving 10,000+ concurrent users at sub-100ms response times. Collaborate directly with clinical researchers and partner institutions.

Drive full deployment lifecycle: architecture, CI/CD, incident response, and capability expansion. Proactively diagnose performance bottlenecks and ship fixes independently.

Translate complex scientific workflows into clean, maintainable production services used by researchers daily.

Built RAG-based clinical AI backend handling 500K+ daily API requests at 99.9% uptime. Redis caching and multi-tenant schema redesign cut query latency 42%.

M.S. thesis: 94.8% performance improvement, 100% retrieval precision, and 86% memory leak elimination on a production clinical AI system.

Reduced transformer inference latency 27% via CPU cache-aligned attention optimization. Published at UTSA ECE 2024; submitted to ACL Rolling Review 2026.

Delivered NLP and ML pipeline contracts for healthcare clients — end-to-end inference systems handling 10,000+ daily predictions, from model training through production deployment.

Built text classification and entity recognition models for clinical data workflows. Managed full project lifecycle including API integration, documentation, and post-deployment support.

Led 8-engineer team building autonomous robot navigation with distributed sensor fusion (LIDAR, IMU, computer vision). Achieved 35% accuracy improvement over 200+ test runs.

Delivered 40% API response time reduction for a platform serving 100K+ daily users via caching and database optimization. Cut peak-load failures 60%.