Miguel Ángel A.

I am a Senior Innovation Engineer on the IT Innovation team at Mutua Madrileña, where I design and develop artificial intelligence systems applied to the insurance sector. With over 8 years of experience in software development —the last few in application architecture— I specialize in building conversational AI agents, multi-agent pipelines, and the cloud infrastructure that sustains them in production.

Among the projects I collaborate on, Dori stands out, a multi-source conversational agent that integrates 6 corporate systems (LeanIX, BMC Discovery, SharePoint, Jira, Confluence, Viva Insights) using LangGraph and intelligent LLM routing: Gemini Flash for quick tasks, Claude Sonnet for complex reasoning. I also participated in T3chi, an Agent-to-Agent orchestrator with end-to-end distributed tracing over Langfuse, and the corporate Tech Radar, a Next.js + Strapi portal that centralizes the organization's technological strategy.

Main stack: Python (FastAPI, LangGraph, LangChain, Chainlit), Azure AKS with ArgoCD for GitOps, Google Cloud (Vertex AI, Discovery Engine), React/Next.js, PostgreSQL, ChromaDB, and Qdrant for data and vector search. Observability with Langfuse and Dynatrace.

I am particularly interested in the frontier between reliable AI systems and their real-world production deployment: context management, tracing, model routing, and the design of agents that function in complex corporate environments.

Available for consulting, development, and innovation projects in AI and cloud, remotely from Madrid.