Applied Engineering
LLMOps, MCP & AI Systems
Running conversational AI in production — where token cost, latency, and reliability matter as much as raw model quality.
LangChain / LangGraphCrewAI / AutoGenMCPCost/token optimization
The day-to-day work at Natura & Co: designing conversational agents that hold up in production, and driving down inference cost through prompt redesign, caching strategies, and model routing. This includes agent frameworks and the Model Context Protocol (MCP) for connecting models to tools and data, plus the architecture decisions that let an AI system scale across a real organization, not just a demo.