How Atelo builds a course for one person.
A look at the real product, feature by feature — the placement that finds your level, the course built around it, the lessons you actually run, and the proof you leave with.
- 01Placement
- 02Generation
- 03Lessons
- 04Challenge
- 05Archaeology
- 06Capstone
- 07Proof
It finds your real level first.
A short adaptive test adjusts to each answer and stops once your level is clear. The result is a skill profile, not a score. Sample learner below: solid on syntax, shaky on predicting what code does.
Sample placement: 11 answers — syntax solid, three of four runtime-prediction questions missed.
Then it builds the course around that.
A full course — modules, a capstone, skill tracking — built around your level, goal, and project, and checked before you ever see it. Here's a real one.
Lessons you run, not just read.
Every lesson is predict-run-repair on a live runtime: predict the output, run the real code, fix what you got wrong. Inline quizzes, a short warm-up drill, and Ada — who gives hints, never the answer. A concept you miss comes back in a later lesson.
Not a screenshot — predict the output and run it yourself:
Each module ends in a build.
Write a plan first, then real code against visible and hidden tests. Ada has your challenge, plan, and tests as context and gives hints only — never the final answer. The module gate opens once every test passes.
Read real production code.
After a module, walk real project code line by line and name the patterns you just learned — control flow, state, error paths. Recognising a concept in code you didn't write is what makes it stick.
Your work becomes one project.
The code you wrote across modules assembles into a capstone that runs on the same real runtime — then you record one architectural decision in your own words. Real assembled work, not a quiz score.
It ends with proof you can verify.
A proof page recomputed server-side from your recorded work and signed, so it can't be faked from a URL. At publish, your capstone re-runs on fresh data it never saw — a hardcoded answer fails. Then it's a link you can share.
One course, built for one person — and proven on real code.
Adaptive placement, a course generated and validated around the gap it finds, predict-run-repair lessons on real code, and proof that re-runs server-side. Early access — every course is $15, built for you, yours to keep.
- 01Placement
- 02Generation
- 03Lessons
- 04Challenge
- 05Archaeology
- 06Capstone
- 07Proof