Full-stack developer who turns business requirements into working AI systems in production. At a brand protection company I build and integrate LLM-powered pipelines (OpenAI, Google Gemini) across REST APIs, databases, targeted scraping, and web-search layers for enterprise clients such as Itaú and Kimberly Clark — with measurable results: a 60% reduction in API token costs and a classification system that cut unnecessary listings by 30% at 90% accuracy. Day-to-day stack is TypeScript, Node.js, MySQL, and MongoDB inside a microservice architecture where stability is critical and error tolerance is low. I care about shipping systems that survive contact with the real world, not model benchmarks.