Responsible for developing robust backend systems that drive clients' businesses through efficient and scalable technological solutions.Flexible Architecture: Development of monolithic and microservices systems, tailored to the specific needs of each project to ensure scalability and efficiency.Powerful APIs: Creation and maintenance of RESTful and GraphQL APIs for efficient service communication, using NodeJS and Java.Monitoring and Performance: Implementation of monitoring systems that improved backend system visibility and performance. Utilized Prometheus and Grafana for continuous system monitoring, while optimizing critical bottlenecks with NewRelic, enhancing user experience.Automation and Agility: Development of CI/CD pipelines that streamlined the software lifecycle, enabling continuous integration of new features and increasing developer productivity.

I developed and implemented an automated system for the isolated recognition of Brazilian Sign Language (LIBRAS) signs using advanced computer vision and deep learning techniques. My work involved extracting skeletal poses from videos and applying various learning models, such as LSTM, ST-GCN, SL-GCN, and Transformers, to capture and analyze gestures over time. To optimize prediction accuracy, I employed an ensemble approach, combining model outputs through a voting mechanism, achieving a superior accuracy rate of 95.3%.Additionally, I contributed to implementing a real-time solution capable of processing videos and returning the most probable sign with high precision via an efficient and robust pipeline. This project underscores the importance of technological innovation in overcoming communication barriers and lays a solid foundation for developing more accessible and effective assistive systems.Key Achievements:Implemented a LIBRAS sign recognition system with high accuracy (95.3%).Utilized advanced deep learning and computer vision techniques.Developed a scalable and robust real-time solution for practical applications.Tools and Technologies:Programming: Python, TensorFlow, KerasDeep Learning Models: LSTM, ST-GCN, SL-GCN, TransformersApplications: Video Processing, Pose RecognitionTechniques: Computer Vision Pipelines

Full-stack SWE at @100Devs. Designing, building, and testing accessible websites and web applications.Skills: HTML · CSS · React · Node.js · Express.js