A stylized classroom scene where a teacher presents maps and diagrams on a large board while a diverse group of students watches attentively, surrounded by books and notes.

A Practical AI Training Plan Across OpenAI, Anthropic, and Microsoft/GitHub

A fast-start learning path from literacy to agents

written by
Mapping the Future: Learning in Motion: an image by

A friend and former colleague reached out and asked for a quick AI training plan he could put to work immediately. I wrote one, then decided to post it here so everyone can use it, adapt it, and improve it.

If this helps, improve it for your context and share it with someone else who is trying to level up.

OpenAI

  • OpenAI Academy — good first stop for broad AI literacy through more advanced engineering-oriented material. OpenAI describes it as covering workshops, discussions, and digital content from foundational AI literacy to advanced integration for engineers. (OpenAI Academy)
  • OpenAI Developers Learn + Quickstart — the fastest path to seeing how the platform works. The Learn hub collects docs, videos, community resources, and cookbook material, while the Quickstart walks through the first API call. (OpenAI Developers)
  • Prompt engineering guides + Cookbook — this is where users should spend real time after the quickstart. OpenAI’s prompt engineering docs cover how to structure instructions, and the Cookbook gives practical recipes and example code. (OpenAI Developers)
  • Agents SDK + Evals — once users understand basic prompting, these are the right next docs for building real workflows, tool use, and quality measurement. (OpenAI Developers)

Anthropic

  • Anthropic AI Fluency + Learn hub — strong starting point for understanding capabilities, limitations, and practical collaboration patterns with AI before users get deep into code. (Anthropic)
  • Build with Claude + Get started with Claude — this is the hands-on path. Anthropic’s learning hub points to the developer docs, and the “Get started” guide walks through a first API call and a simple web-search assistant. (Anthropic)
  • Prompting best practices — Anthropic’s core prompt guide is worth reading carefully because it is very practical and very builder-focused. (Claude API Docs)
  • Building with the Claude API + Tool use + MCP — this is the next layer after basics: structured course material, function/tool calling, and MCP for connecting models to external systems. (Anthropic)

Microsoft / GitHub

  • Microsoft Generative AI for Beginners — probably the best Microsoft starting repo for modern GenAI. If users want a broader AI foundation first, Microsoft’s AI for Beginners is the companion option. (GitHub)
  • GitHub Models Quickstart + GitHub Models docs — this is especially useful because GitHub says you can run models with your GitHub credentials, then prototype, optimize, evaluate, and store prompts right in the repo workflow. That makes it a very practical “learn by doing” path. (GitHub Docs)
  • GitHub Copilot prompt engineering — use the GitHub Docs page together with Microsoft Learn’s prompt-engineering module and Copilot learning path. This is great for someone who already lives in engineering and operations workflows. (GitHub Docs)
  • AI Agents for Beginners + Semantic Kernel — after fundamentals, this is the best Microsoft path into agentic systems. Semantic Kernel is Microsoft’s model-agnostic SDK for building AI agents in C#, Python, or Java, and the quickstart plus MicrosoftLearning repo make it concrete. (GitHub)

The Order

  1. Start broad: OpenAI Academy, Anthropic AI Fluency, and Microsoft Generative AI for Beginners. (OpenAI Academy)
  2. Do one vendor quickstart end to end: OpenAI Developer Quickstart or Anthropic Get Started. (OpenAI Developers)
  3. Use GitHub Models next: it lowers setup friction and helps users compare models and prompts in a familiar GitHub workflow. (GitHub Docs)
  4. Study prompting seriously: OpenAI prompt guides, Anthropic prompting best practices, and GitHub Copilot prompt engineering. (OpenAI Developers)
  5. Then move into agents and enterprise orchestration: OpenAI Agents/Evals, Anthropic tool use/MCP, and Microsoft Semantic Kernel. (OpenAI Developers)