Skip to main content
⚡
CodeMentor AIAll tracks
🤖L3 · 7 modules · 101 lessons

AI Engineering with Python

Build with Claude & LLM APIs like the US dev community does

100 lessons across 6 modules: API fundamentals, tool use & function calling, RAG, agent loops, production AI, and frontier topics (vision, prompt caching, compaction, Files/Skills/Batches APIs, voice agents, multi-agent orchestration).

Browse all 101 lessons →Try the first lesson — no signup
101
lessons
7
modules
3
capstone missions
18
languages
What you'll learn

From day one

  • ✓Anthropic SDK setup, model selection, adaptive thinking
  • ✓Tool use via SDK runner — let Claude call your Python functions
  • ✓RAG: chunking, embeddings, BM25 + dense retrieval, reranking
  • ✓Agent loops: planner → workers → synthesizer patterns
  • ✓Prompt caching for 90% cost reduction at scale
  • ✓Production: streaming, compaction, Files API, Batches API
  • ✓Vision inputs, voice agents, multi-modal apps
  • ✓Eval-driven prompt iteration with versioned golden sets
  • ✓Capstone: end-to-end customer-support agent with all production constraints
Curriculum

7 modules · 101 interactive lessons

#1Module 0 · Before you start
#2Module 1 · LLM API Fundamentals
#3Module 2 · Tool Use & Function Calling
#4Module 3 · RAG (Retrieval-Augmented Generation)
#5Module 4 · Agent Loops & Workflows
#6Module 5 · Production AI Engineering
#7Module 6 · Frontier AI Engineering
See all 101 lessons →

Frequently asked

Is AI Engineering different from Machine Learning?

Yes. ML builds the model. AI Engineering ships it — Claude/GPT APIs, RAG, tool use, eval suites, prompt versioning, production observability. This track is the ship-it half.

Do I need to know Python before starting?

You should be comfortable with Foundations or equivalent (functions, classes, async basics, requests). The first 3 lessons cover the absolute prereqs.

Which LLM does the track teach against?

Claude (Anthropic) as the primary; OpenAI patterns where they differ. The eval + RAG patterns translate to any frontier model.

Will I be able to ship something to production after this?

Yes — the capstone is a Claude-powered tool-using chatbot with RAG, a 30-PR golden eval suite, and observability. Deploys to Fly.io.

How fresh is the content for 2026?

Updated this year. Covers tool use, prompt caching, Files API, Skills, compaction, voice agents, multi-agent orchestration — the late-2025/2026 stack.

First 15 lessons free, no signup

All 101 lessons unlock for $12/month or $89/year. 7-day free trial — card required, cancel in 1 click anytime in Settings.

See pricing