Applied AI projects — building with models and APIs

Junior AI developer, technical freelancer, product builder — I build a working tool on top of an AI model and hand it to a real user.

Level 2 — technical (Python)

You finished Machine Learning Foundations, or you already code. You want something a user can open, not a notebook that only you can run.

What you'll be able to do

• Call a model API from Python and handle errors, limits and cost
• Build a question-answering assistant over a set of your own documents
• Ship a small web app a non-technical user can use
• Write an evaluation that shows whether the tool actually helps
• Document limits, failure modes and what a human must still check

Modules

• M1 — Calling a model from code — Requests, keys, errors, retries and cost per call.
• M2 — Assistants over your own documents — Retrieval, chunking and answers that cite the source.
• M3 — Shipping an app — A screen, an input, a result — for a non-technical user.
• M4 — Evaluating a real tool — Test sets, user trials and the honest verdict.
• M5 — Handover — Documentation, cost, maintenance and the limits page.

Capstone project

A tool in someone else’s hands — Build an AI tool for a real person or organisation, deploy it, have them use it for two weeks, evaluate it against a test set and user feedback, and hand it over with documentation and a budget.

What you earn

• A deployed tool with a public link
• A written evaluation and limits page
• LES Institute certificate — Applied AI Projects

Requirements

• Machine Learning Foundations, or solid Python
• A GitHub account
• A computer, not only a phone