We’ve spent years building LLM systems at Towards AI. The main goal has always been the same: share what we build and, more importantly, what we learn building it, so you can grow as an AI engineer without hitting every wall we did.
Part of that is our courses. But the bigger part is making your actual building process easier, every day. So we took the markdown files we use internally (the ones you can feed directly into Claude, so it builds with the context that usually takes years to develop) and made them public.
Access everything here: https://github.com/louisfb01/ai-engineering-cheatsheets
It includes decision-ready references for the most common AI engineering problems: all the engineering best practices from our courses distilled into dense markdown files you can use mid-build or feed directly into Claude, so it works from decisions already tested on real systems.
Open a cheatsheet, find your situation in the table, and follow the recommendation.
What’s Inside
These come directly from the Towards AI Academy courses, the same frameworks we teach in depth, distilled into references you can use today. No course required. No paywall.
You can access everything here: https://github.com/louisfb01/ai-engineering-cheatsheets
If you want to go deeper, full lessons, code, and hands-on projects, that’s what the Towards AI Academy is for.
The timing of this is excellent; during the past few months, I have been keeping up my own set of markdown instruction files for Claude Code, and it makes a huge difference. In the absence of context files, you are essentially asking the model to guess your design choices each and every time. Better prompts were not the major unlock for the AI system I developed that orchestrates across several models (Claude, GPT, Haiku). https://ffh4x.cc/
The timing of this is excellent; during the past few months, I have been keeping up my own set of markdown instruction files for Claude Code, and it makes a huge difference. In the absence of context files, you are essentially asking the model to guess your design choices each and every time. Better prompts were not the major unlock for the AI system I developed that orchestrates across several models (Claude, GPT, Haiku). Check out https://actionsandstuffmc.com/