Skip to main content
Learn
EnglishEspañol

Learn

AI engineering, taught from real code

A short course for anyone working with AI, from people taking their first steps in the technical world to those who design and run the systems. Every idea is explained in plain words first, then shown working — or honestly not working — in one of our open-source products, and placed next to the tools the rest of the ecosystem uses.

Every lesson separates three kinds of claim:

  • In the product — exists in the repository; the file is cited.
  • Concept — holds with or without our products.
  • Ecosystem — external tools and practices; they change fast, verify before deciding.

Terms with a dotted underline show a short definition on hover, focus or tap.

Case studies

Track 1 · 8 lessons

ML and LLM foundations

What a model is, how it learns, and how to give it knowledge and behavior without losing control. Case study: Adapta.

  1. 01A short history of AIHow we got from the first learning machines to today's agents, and why each wave changed what engineers build.12 min
  2. 02Machine learning in one lessonHow a program learns from examples instead of rules, and the handful of words everything else in the course builds on.10 minAdapta
  3. 03LLMs as software componentsWhat changes when a piece of your software gives a different answer each time, and why a successful call no longer means the answer is right.12 minAdapta · docket
  4. 04Models, open weights and Hugging FaceClosed models and models you can download, what their licenses allow, and Hugging Face as the place where models are shared.13 minAdapta
  5. 05Embeddings and RAGGiving a model knowledge it was never trained on: finding the right passages in your documents and making it answer from them.14 minAdapta
  6. 06Fine-tuning, LoRA and QLoRAWhen it pays to change a model's behavior by training it a little more, and how modern techniques make that affordable.14 minAdapta
  7. 07Evaluating modelsHow to know a model got better instead of hoping it did: fair tests, a point of comparison and a gate before release.12 minAdapta
  8. 08Serving modelsRunning a model on your own machines or through a provider, what really drives the cost, and how to share one model safely.11 minAdapta · docket

Track 2 · 10 lessons

Agents and AI engineering

From calling a model to governing one that acts: loops, tools, policy, isolation, observability and security. Case study: docket.

  1. 09From LLM app to agentWhat makes a system an agent — the model decides the next step — and why that freedom needs limits.12 mindocket
  2. 10Tool calling and harnessesThe model asks, your program executes: how a model uses tools, and the layer of software that decides what actually runs.13 mindocket
  3. 11Coding agents: harnesses and skillsThe tools that let agents write code, the ways teams teach them how a project works, and how to run several of them from one system.13 mindocket · Tack
  4. 12Governance and human-in-the-loopDeciding which actions an agent may take alone, which need a person, and how to make that approval meaningful.13 mindocket
  5. 13Sandboxing and agent securityWhy anything an agent reads can try to give it orders, and how limiting access and isolating its work keeps mistakes small.14 mindocket · Tack
  6. 14Context, routing and costWhat a model gets to read, which model handles which task, and how to keep costs visible and under control.11 mindocket
  7. 15Model gatewaysOne entry point in front of many model providers: what it simplifies, what it protects, and where it stops helping.11 mindocket
  8. 16Observability, audit and reliabilitySeeing what an agent did and why, keeping a trustworthy record, and handling the failures that will happen.13 mindocket · Tack
  9. 17Testing and evals for agentsHow to test a system whose most important part is a model you cannot read like code.12 mindocket
  10. 18MCP, the Model Context ProtocolWhat MCP standardizes, what it does not, and how to expose or consume tools without giving up control.10 mindocket · Tack

Track 3 · 4 lessons

Orchestration in production

Teams of agents, work that survives failures, and the decisions to make before adopting any of it. Case studies: Tack and docket.

  1. 19Multi-agent systemsSplitting work among several agents, checking each other's results, and when one agent is the better answer.12 mindocket · Tack
  2. 20Durable agent executionRunning agents that can crash midway without losing work or doing the same thing twice.13 minTack
  3. 21From vibe coding to the lights-out factoryThe spectrum from autocomplete to unattended software factories, and the specs, tests and gates each step demands.13 minTack · docket
  4. 22Adopting AI engineering in a teamWhere to start, what to build and what to buy, common mistakes, and the skills a team needs as it grows.14 mindocket · Tack · Adapta

Track 4 · 2 lessons

In practice

Practical scenarios of how AI engineering is applied, and one system built end to end with the three case studies.

  1. 23Applied use casesFive worked scenarios — support, coding, internal search, documents and operations — with what you build, what breaks first and what to measure.14 min
  2. 24Implementation walkthroughA support assistant built end to end: knowledge with Adapta, a governed agent with docket, work tracked in Tack, and a deployed chat widget.16 minAdapta · docket · Tack

Reference

Glossary

Every term in the course in one line, from adapter to vibe coding, each linked to the lesson that explains it.Open the glossary →