AI Ustoz UZ: learn artificial intelligence in Uzbek

Coding path

A deeper path for gentle beginners and vibe coding: Python, the cycle of programming with artificial intelligence, and code exercises. Each level opens after the previous one is finished. The exercises run right in the browser.

Open in the app

Interactive mode: search, progress and Python exercises.

Beginner level

Python and artificial intelligence

Learn the basics of Python and see, through a small “agent” example, how programs that use artificial intelligence are built.

  1. How programs work with artificial intelligence8 min
  2. Python basics: print, variable, f-string, list and loop14 min
  3. Dictionary and JSON: structured data12 min
  4. Functions: code you can reuse12 min
  5. A simple “agent”: a program that picks a tool14 min

Project: “Mini assistant”

Write a small Python program: a dictionary with a few questions and answers, a function that returns the matching answer, and an “I do not know” case for an unknown question. Test it with 3 questions.

Intermediate level

the vibe coding workflow

Learn the cycle of programming with artificial intelligence: write a spec, ask for code, read it, test it and fix the error. The main principle: do not trust code written by artificial intelligence blindly.

  1. Vibe coding and a precise spec12 min
  2. Asking artificial intelligence for code and reading it12 min
  3. Testing code and finding a bug14 min
  4. Reading an error message and fixing it12 min
  5. A small project: a task list15 min

Project: “Task list” program

Write a small program that adds tasks, marks them done and shows the ones left. First write the spec (input, output, 3 examples), then write the code yourself or ask a chat tool for it, and then read, test and fix it.

Advanced level

API, tool calls and the idea of RAG

Understand how a program talks to artificial intelligence services: a structured request and response, tool calls and the idea of answering from a found document. All exercises are simplified for learning: there is no connection to a real service.

  1. API: a structured request and response14 min
  2. Tool calls: the model chooses, the program runs14 min
  3. The idea of RAG: find first, then answer15 min

Project: Simplified “answer from a document” example

Make a list of 3–4 short texts. Find the text that best fits a question, put it into the request text and build a ready prompt. You do not need to connect to a real model: you check the prompt yourself.