Rafael Fonsêca

Creation of Artificial Intelligence models using supervised and unsupervised learning in the field of NLP (Natural Language Processing), as well as the implementation of the system, from the development stage of the system architecture to the implementation of the system in Desktop and Web applications. Collaborated on stages of the systems development lifecycle from requirement gathering to production releases. Tested and deployed scalable and highly available software products. Some of the techniques/technologies used include:

• Clustering, Word Embeddings, nGrams, LLama, BERT (and some variations like RoBERTa), BigBird, prompt engineering.
• Image preprocessing for OCR. Qualitative evaluation and automatic correction of the extracted texts.
• Python, FastAPI for backend development, React for frontend development, and Docker for containerization.

• Gathered and organized information for research purposes.
• Utilized statistical packages to analyze data and create visualizations.

Work carried out to implement reinforcement learning for humanoid agents using a custom OpenAI Gym environment to support the training of agents running in a robot soccer simulator. The main focus of the project is on kick policy decision-making. This involved training agents using reinforcement learning to develop decision policies. To achieve this, it was necessary to create a proxy for data collection from the server running the simulation and to create the environment for agent training.

Technologies used: Python, OpenAI Gym, C++.

Created APIs with NodeJS (Backend) and developed mobile applications using React Native. Reviewed code, debugged problems, and corrected issues. Applied agile methodologies including Scrum and Kanban to manage project workflows and deliverables.

Technologies used: NodeJS, React Native.

Creation of software for web scraping (information extraction) using Python and the Selenium framework. Also participated in bug tracking, web application development, and testing of existing web applications.

Technologies used: Python, Selenium, HTML, CSS, JavaScript, NodeJS.

Requirement gathering, software architecture and implementation of automations for invoice issuance and data collection, reducing the manual work of company employees.

Technologies used: Python, Selenium.