At Altion, I build AI assistants for scheduling and customer service, grounding answers in client content through RAG pipelines with embeddings and vector search. I also develop conversational agents with LangChain and LangGraph, and an MCP server that exposes internal tools and data to models.
I connect LLMs to REST APIs and systems such as scheduling, CRM, and WhatsApp, and add validation, scope restrictions, and human escalation when confidence is low. I tune chunking, context, caching, and model selection to reduce token use and response time without losing quality.
At Cognition, I build enterprise automations across ERPs, web portals, and databases, and document requirements and testing through PDD, SDD, and UAT. Earlier, at Cia. Hering and Grupo SOMA, I worked on business-process automation, from identifying candidates and building bots to supporting adoption and daily operations.
