Engineering courses for people who build systems without writing much code. You work through them with the AI you already use. There is nothing to install.
openAgent engineering for people who build systems without writing code
You have been using an AI coding agent for a few weeks and hit the point where it stops feeling like magic and starts producing confident garbage. This is the engineering underneath that: what to ask for, what it can see, whether it actually did the thing, and how to prove it.
None of these have a date yet. Pick the one you would take and it gets built next.
Evals for systems that do not give the same answer twice
"How do you know it works?" is the question every AI engineering interview turns on, and most answers are a number with no derivation behind it. Golden sets, rubrics, pass bars, sampling, and judges that can be fooled. Forward Deployed spends one unit here. This is the whole thing.
MCP servers, and the agents that use them
You take Forward Deployed through an MCP server. This one has you build one: designing tools an agent can actually use, deciding what to expose and what to withhold, auth, and the difference between a tool that answers and a tool that does.
Memory, staleness, and retrieval across agents
Agents forget between runs, and the fixes are where most agent systems quietly go wrong. Schema design for things a model will read, indexes that do not bloat, retrieval that surfaces the current answer rather than a confident stale one, and memory several agents write to at once.
One email when it opens. Nothing else, ever.