All work
Case study

Teaching AI safety and practical AI to students and educators

A year of hands-on AI teaching in Mexico after Microsoft: AI safety and digital citizenship for students aged 13 to 18, and workshops for educators and faculty.

Period
2025 to 2026
Organization
Independent, Playa del Carmen, Mexico
My role
Consultant, teacher and builder

Context

After leaving Microsoft in July 2025 I spent a year on hands-on AI work outside the Microsoft stack: teaching, workshops and prototypes, to understand what the different models and tools are actually good for before advising anyone on them.

What I did

  • Taught AI safety and digital citizenship to students aged 13 to 18 at a summer school in Mexico: how the tools work, where they fail, and how to think critically about what they produce.
  • Ran hands-on AI workshops for educators and faculty covering large language model fundamentals, prompting, personalization with custom instructions and memory, responsible use, and content creation for teaching.
  • Ran workshops for organizations and solo entrepreneurs on brand voice, websites, chatbots grounded in their own material, and when to use which tool.
  • Built education prototypes: a learning management system with a backend, admin and user web interfaces and a native iOS app, plus smaller tools such as a local live translator.

Scale and outcome

  • Small groups, taught in person. I do not publish participant counts.
  • A workshop format that starts with a workflow rather than a tool, which I still use.
  • Working knowledge of models and tools from OpenAI, Anthropic, Google, Perplexity and local inference, which now informs Copilot and AI adoption work with districts.

What it means for a district

A first educator workshop should start with one job teachers already do, define what the tool may and may not use, and put the teacher’s review back in the loop. Demos impress; a reviewed workflow is what gets adopted.

Last reviewed October 27, 2026.