Every lab runs on your own computer from Fleet.

AI engineering · Part 01Talk to a model

Run a model, call it from code, compare models with numbers, and give one tools.

• 10Run a model on your laptopbeginner · 15 min · 3 stepsNot started
• 11Call a hosted model and read the billbeginner · 25 min · 4 stepsNot started
• 12Pick a model with numbersbeginner · 30 min · 4 stepsNot started
• 13Give a model tools: the loop from scratchintermediate · 40 min · 4 stepsNot started
• 14Get data back, not proseintermediate · 30 min · 4 stepsNot started

AI engineering · Part 02Give it knowledge

Search by meaning, answer from your own documents, measure it, and fine-tune a small model.

• 20Search by meaningbeginner · 30 min · 4 stepsNot started
• 21Answer from your documents, with sourcesintermediate · 45 min · 4 stepsNot started
• 22Measure retrieval before you tune itintermediate · 40 min · 4 stepsNot started
• 23Build a dataset and a baseline to beatintermediate · 45 min · 4 stepsNot started
• 24Fine-tune a small open modeladvanced · 90 min · 4 stepsNot started

AI engineering · Part 03Build the harness

The parts around the model: permissions, budgets, tool servers, and checks on generated code.

• 30A model plus a harnessintermediate · 45 min · 4 stepsNot started
• 31Put a coding agent on railsbeginner · 30 min · 4 stepsNot started
• 32Build a tool server any assistant can useintermediate · 45 min · 4 stepsNot started
• 33Make AI-written code safe to mergeintermediate · 40 min · 4 stepsNot started
• 34Choose and run open modelsintermediate · 35 min · 4 stepsNot started

AI engineering · Part 04Make it reliable

Grading, traces, memory, flows that survive a restart, parallel work, and safe self-improvement.

• 40Grade with a judge you have checkedintermediate · 50 min · 4 stepsNot started
• 41See every step: tracesintermediate · 40 min · 4 stepsNot started
• 42Memory that outlives the conversationintermediate · 40 min · 4 stepsNot started
• 43Survive a restart, wait for a personadvanced · 60 min · 4 stepsNot started
• 44Many workers, one dispatcheradvanced · 45 min · 4 stepsNot started
• 45Let it improve itself, and prove itadvanced · 60 min · 4 stepsNot started

AI engineering · Part 05Deliver to a customer

Choose the right problem, attack your own work, control every tool call, ship it, and hand it over.

• 50Find the use case worth buildingbeginner · 45 min · 4 stepsNot started
• 51Attack your own assistantadvanced · 60 min · 4 stepsNot started
• 52One door for every tool, with rulesadvanced · 50 min · 4 stepsNot started
• 53Ship it and know the billintermediate · 50 min · 5 stepsNot started
• 54Capstone: the claims deskadvanced · 180 min · 4 stepsNot started

Flisk CLI · Part 01Foundations

Get oriented, confirm your CLI, and make a project that Flisk can understand.

• 01Is Flisk installed?beginner · 5 min · 2 stepsNot started
• 02Scaffold a projectbeginner · 8 min · 2 stepsNot started
• 03Add a backend resourcebeginner · 10 min · 2 stepsNot started

Flisk CLI · Part 02Project structure

Learn how projects, stacks, and resources fit together.

• 02Scaffold a projectbeginner · 8 min · 2 stepsNot started
• 03Add a backend resourcebeginner · 10 min · 2 stepsNot started

Flisk CLI · Part 03First service

Create a backend resource and inspect the generated contract.

• 03Add a backend resourcebeginner · 10 min · 2 stepsNot started