Node.Js Developer

Developed an intelligent inventory management solution for DASA during the FIAP Annual Challenge. The strategic goal was to mitigate human error and drastically accelerate the check-in/check-out process by replacing manual workflows with high-speed automation. Due to its technical impact, the project ranked among the Top 10 Software Engineering projects, leading to a showcase at the FIAPNEXT innovation event and academic honors (exam waivers). Key achievements and responsibilities: Hybrid AI Engineering: Engineered a dual-layer AI system for item recognition. I implemented the capture and initial visual detection layer on the front-end using React and TypeScript. Computer Vision: Developed and trained a YOLO (You Only Look Once) model, enabling ultra-fast initial identification and allowing end-users to contribute to continuous model training. Back-end Orchestration: Built the core business logic and RESTful integrations using the Spring Framework (Java). I structured a sophisticated pipeline where images, after initial YOLO detection, were processed and sent to the Google Gemini API for a second layer of deep validation and data cross-referencing. Infrastructure & DevOps: Containerized the entire ecosystem using Docker for local deployment, ensuring an isolated, stable, and easily reproducible environment for the integration between AI services, the Java API, and the user interface. Tech Stack: Java, Spring Framework, TypeScript, React, YOLO (You Only Look Once), Google Gemini API, Docker.