Carlos Armando Calderon Cortes

Technology leader and solutions architect with over 20 years in enterprise platforms, modernization and engineering leadership. At Quantus Information Technology, I lead more than four engineers and collaborate with a Product Manager. My practical experience in AI agents, multi-agent orchestration, Model Context Protocol (MCP) and LLM applications spans approximately two years.

I direct Codex agent workflows with scoped tasks, repository context, clear ownership, independent review and verifiable acceptance criteria. I improve the agent harness through instructions, tool and data contracts, concurrency controls and evidence-based validation. I led the Q Agro AI assistant with authorized company context, structured outputs, validation, quotas and idempotency, and MCP integrations with Firebase Auth and Firestore for auditable agent reports.

For Ráfaga, I iterated grounded AI prompts and evaluations: an eight-case baseline improved from 3/8 to 8/8, followed by 24/24 across three qualification runs. These are evaluation results, not production reliability claims. My work includes JSON Schema validation, source checks, rejection when evidence is insufficient, cost and latency limits, observability and controlled production authorization.

At Globant, I led modernization for Empower and Santander México, including legacy backends and mobile banking, and managed up to 10 people. At Zucarmex, I led enterprise ERP architecture and delivery; earlier I directed biometric platforms and APIs at Biometría Aplicada and banking systems at Prudential / Actinver.

Core strengths: enterprise and solutions architecture, IT strategy, engineering leadership, Java / Spring Boot, TypeScript, React / Next.js, Python, Node.js, GCP / Firebase, APIs, CI/CD and automated testing. TypeScript: approximately five years; Python, Node.js, AI agents and MCP: approximately two years. Seeking remote work from Mexico, or onsite/hybrid roles in Sinaloa or Mexico City.