Miguel Escribano
The future of work, AI, Environment, Business, and whatever I’m learning in public. Written as I go.
Recent Posts
Building an AI agent for indoor air quality: architecture, MCP tools, and a human gate
Anne is an agentic IAQ consultant running inside an AI coding assistant. Live sensor readings, deterministic scoring outside the language model, and a human gate before recommendations.
Notes
AI value = (Productivity × Localization) − (Technical Debt × Moral Debt)
Technical debt has a moral twin. A formula for when AI work produces real value.
Marketing engineers, business engineers
Marketers and business developers are building their own tools. I opened a GitHub account in 2025 and built MCP and skills before the official CRM integration.
We're hiring. Bring your own AI agents.
Employers now expect you to build an AI stack that helps you scale.
Resources
fiware-platform-ngsi-v2-mcp
MCP server that connects AI assistants to FIWARE NGSI-v2 platforms: Context Broker entities, historical data, CEP rules, and IoT device registration.
openair-ai-kit
MCP server and IDE plugin that turns air quality data into openair charts from inside an AI workflow.