Filipe Frozza

Led architecture and delivery of AI-powered marketing automation features, building Retrieval-Augmented Generation (RAG) pipelines to centralize AI memory and orchestrate LLM-driven workflows. Integrated OpenAI GPT-4 Turbo and Anthropic Claude 3 APIs, built a pgvector/Pinecone vector search system, and mentored a cross-functional team of 5+ engineers to improve reliability and delivery velocity.

Designed and delivered scalable fintech SaaS modules for DB1 Finance Hub, including multi-tenant account management and transaction orchestration. Integrated Stripe and Plaid, embedded LLM-based support chat/Q&A, and refactored REST/GraphQL APIs with contract testing and versioning, reducing integration defects by 30%.

Engineered HIPAA-compliant patient data ingestion and eligibility verification workflows for the Lean Health Portal by integrating FHIR APIs and HL7 interfaces for EHR interoperability. Built Kubernetes-based, event-driven microservices and implemented OpenAPI documentation and stricter payload validation to reduce integration errors and support secure audit trails and data access controls.

Designed and delivered scalable e-commerce modules for the DLV Merchant Platform, supporting thousands of concurrent users through containerized microservices. Migrated from a legacy monolith, implemented catalog search/filtering with PostgreSQL and Redis, and introduced automated testing pipelines with Jest, Pytest, and JUnit 5 to reduce production bugs by 25%.

Developed high-traffic exam registration and result management systems supporting 50,000+ concurrent users, optimizing SQL queries to improve responsiveness. Built REST APIs and React 16 components, and improved deployment reliability with CI/CD automation using Jenkins and Docker, reducing page load times by 40%.

Delivered enhancements for the Parce Loyalty Platform, including migration from legacy PHP and AngularJS modules to Java/Spring Boot and React. Implemented authentication/authorization with RBAC and built automated test suites with JUnit and Selenium, increasing test coverage and reducing manual QA effort by 30%.

Maintained and enhanced academic management systems using Java/Spring and legacy PHP, supporting scheduling, grade reporting, and student data management. Upgraded and refactored Python import/export utilities and built custom reporting dashboards with JSP/JavaScript, optimizing SQL Server queries and reducing server load during peak registration periods.

Earned a Bachelor's degree in Computer Science from ULBRA between 2018 and 2022.