At SensorEng, I refactored a legacy monolith into a modular architecture, reducing the application’s initial bundle size by 85%. I also developed features for real-time industrial monitoring that brought together continuous video, AI model inference, IoT sensor telemetry, and georeferenced maps.
At Avante Sistemas, I developed and maintained features for business ERP systems and helped investigate and correct code and database issues. As an intern on Software de Gestão de Crédito, I developed document-reading and data-extraction services using computer vision and OCR, supporting recognition and classification of more than 20 document types.
For Russian Mastery, I built a full-stack language-learning platform with an SM-2 spaced-repetition algorithm, a REST API, Redis caching, and PostgreSQL persistence. My computer vision capstone focused on real-time pose estimation and tracking for boxing; a custom-model detection approach reduced total processing time by 83.6%.

