Gabriel Maestre Costa
Backend and full-stack developer
Experience building and evolving production systems, APIs, integrations, and process automation, mostly for financial solutions.
The work covers PHP, Laravel, Java, Spring Boot, Node.js, PostgreSQL, MySQL, Angular, and TypeScript, along with REST and SOAP, XML and JSON, legacy systems, multi-tenant architecture, and asynchronous processing. Delivery uses GitLab, CI/CD, Docker, Composer, Maven, and SonarQube.
B.Sc. in Computer Science, with experience in AI, reinforcement learning, and multi-agent systems.
Ponta Grossa, Paraná, Brazil
app/api/profile/route.ts
Experience
Fácil Tecnologia
Builds and evolves production systems for payroll-deducted loans, financial operations, collection, and credit recovery. ConsigFACIL is a multi-tenant payroll-loan platform in PHP and PostgreSQL. ProFACIL is a credit-recovery platform, with a Java/Spring Boot API and an Angular/TypeScript interface.
AquiCob Soluções de Cobrança
Built collection modules in PHP and MySQL for agreements, interest, penalties, discounts, and collection operations, and bank integrations over SOAP and REST.
Moonrock Soluções Tecnológicas (contract)
Built RESTful APIs in Laravel/PHP and interfaces in React/Next.js, with PostgreSQL, PHP queues, Docker, and CI/CD.
Projects
Fig. 1 Scientific Initiation, Intelligent Agents
Trained a DQN agent to drive in a simulated city. The control problem was set up in CARLA, the training loop ran in Python with TensorFlow and Keras-RL, and the agent was validated experimentally as a deep network.
Fig. 2 Scientific Initiation, Traffic Simulation and AI
Compared CARLA, SUMO, and Traffic3D before later experiments committed to one of them. The choice rested on realism, behavioral fidelity, computational cost, and how directly each simulator connects to a reinforcement-learning stack.
Fig. 3 MASPY / Thesis
Compared a multi-agent learner with a monolithic one on the same intersection. MASPY traffic-light agents learn with Q-learning inside a BDI environment model; a tabular SARSA baseline trains on an equivalent environment, with shared transitions, reward, and exploration.
github.com/Dev-Maestre/Smart-Queue-Agents-W-Maspy
Education
B.Sc. Computer Science, UTFPR, 2021–2026