At Akad Seguros, I developed an AI-agent solution for claims that translated process rules and decision points into a system that reached 85% accuracy in calculating losses for transport claims. I also developed AI solutions for underwriting and credit risk, from experimentation and validation through deployment and monitoring.
At Targit Brasil, I led the technical evaluation and selection of AI technologies for client needs and guided development of a Text-to-SQL solution using LLMs, RAG, and agents. I structured AI solutions as APIs and developed backend components for system integration.
At CEIA, I developed AI solutions for public and private organizations, including NLP systems for document analysis and information extraction. This work included a system used by TCM-GO to analyze documents submitted by municipalities in Goiás, as well as a multi-agent travel concierge for Bancorbrás.
