At Lumen Technologies, I developed and enhanced backend applications using Java, Spring Boot, SQL, REST APIs, and Kafka. I implemented Security Deposit functionality within an event-driven architecture and validated services across application layers.
I’m building CodeArchaeologist, a developer tool that analyzes repositories to recommend whether code can be safely deleted, preserved, quarantined, or reviewed. Its analysis combines static analysis, Git history, test coverage, and semantic reasoning to produce explainable recommendations.
I’m also architecting DevSquad AI, a multi-agent framework that turns a product goal into a coordinated workflow for producing tested, versioned code. It includes real-time agent observability and a project-memory layer using PostgreSQL and pgvector.
For my sign language recognition project, I implemented a TensorFlow CNN that achieved 94.82% test accuracy and used OpenCV in a pipeline that converts recognized signs into speech. The related research was published in IEEE Xplore.
