At Independent / Freelance, I integrated OpenAI and Anthropic LLM APIs into SaaS products, reducing manual workload by ~35%. I also built AI-driven data-analysis pipelines that made insight generation ~40% faster.
I cut AI API and inference costs by ~25% while maintaining output quality, and reduced AI-output error rates by ~25% with monitoring and observability. Reusable AI integration layers helped accelerate feature iteration by ~30%.
Previously, at Via Nova, I improved Go and Python backend performance by 30–50% and delivered full-stack applications with React, TypeScript, JavaScript, Node.js, and PostgreSQL. I also worked on Solidity/EVM integrations for production Web3 applications and conducted ontology and knowledge-representation research at Pontifícia Universidade Católica do Paraná (PUCPR).
