Python → AI track · free while in beta

Tutorials don't get you hired. Built things do.

Build three job-realistic AI projects — a data pipeline, an ML service, a RAG chatbot — with an AI mentor reviewing every milestone like a senior engineer reviews a pull request.

Start building — it's free See the track

mentor review · milestone 3/3

verdict: pass

Solid pipeline. Your upsert on (city, timestamp) makes re-runs safe — that's the thing most production incidents trace back to.

strengths

– raw responses saved before transforming, replayable cleaning

issues

– your log swallows the traceback, so cron can't see failures

hints

– look at log.exception, then re-raise — let it fail loudly

Learning that works like the job

Build on your machine

Real editor, real terminal, real Git history — the exact workflow of the job, not a browser sandbox.

Quizzes gate the concepts

Short, tricky checks after each lesson. You can't cargo-cult your way forward.

An AI senior engineer reviews your code

Paste your code and output at each milestone. Get a pass/retry verdict with strengths, issues, and hints — never the solution.

The Python → AI track

Three projects that build on each other: the data you pipeline in Project 1 trains your model in Project 2; its docs feed your RAG chatbot in Project 3. Finish with a portfolio that tells one story.

beginner ~8h

Build a Weather Data Pipeline

Fetch real weather data from a public API, clean it with pandas, load it into SQLite, and automate the whole thing — the exact pipeline pattern used in real data teams.

PythonrequestspandasSQLitecron

intermediate ~10h

Train and Serve a Weather Prediction Model

Use the weather data your Project 1 pipeline collected to train a model that predicts tomorrow's temperature — with honest baselines, leakage-free features, real evaluation, and a live FastAPI endpoint.

Pythonpandasscikit-learnjoblibFastAPISQLite

advanced ~12h

Build a RAG Chatbot Over Your Own Docs

Build a Retrieval-Augmented Generation chatbot in Python that answers questions over your own project docs — embeddings from scratch, cosine-similarity search, grounded prompts with citations, and a real eval set. The exact architecture behind every "chat with your docs" product.

Pythonanthropicsentence-transformersnumpyClaude