a real ai program
A chatbot and a classification pipeline calling a live model API — your first program that thinks.
The free course for people who have never touched AI — and don’t know where to start. If you can use a web browser, you can start today.
$ git clone github.com/crackcode09/zero-to-ai-systems-engineer
$ cd zero-to-ai-systems-engineer/00-orientation
# no setup. no install. just open the README.Twelve modules, in strict order — each one builds on the last. Skipping ahead is the #1 reason beginners quit. The build modules are where a portfolio piece lands on your GitHub.
explain ai, ml & llms in plain english; use ai hands-on
the terminal, git, github and vs code
real programs: data, logic, files, packages
speak json, call apis, query sql
training vs inference, weights — the honest version
tokens, context windows, why hallucination happens
reliable, structured outputs on demand
a chatbot + an ai classification pipeline
doc-q&a from scratch in 80 lines
the agent loop, and what mcp really is
measure ai quality; reason about risk like a pro
ship a hardened fastapi service
the project that proves all of it
You don’t end with a certificate. You end with things that run — four portfolio projects and a capstone, all on your GitHub.
A chatbot and a classification pipeline calling a live model API — your first program that thinks.
Retrieval-augmented Q&A built from scratch in ~80 lines. It answers from YOUR files, not the internet.
The agent loop, tools and MCP — an assistant that does things, not just talks.
A hardened FastAPI service: rate limits, errors handled, ready to deploy.
Pair this with the domain you already know — healthcare, manufacturing, finance, teaching — and you become genuinely rare.
Modern AI is an incredibly capable but unreliable component. An AI Systems Engineer builds systems that harness the capability while containing the unreliability.
Every module maps to authoritative material. This repo is the beginner on-ramp those resources assume you already have.
Arithmetic. That’s it.
No — the AI runs on cloud servers. A browser is enough.
~$5 in API credits for the whole course (modules 07+). Everything else is free.
It builds a junior-level portfolio of 4 projects + a capstone. Paired with what you already know, that’s rare.
The best time to start learning AI was three years ago. The second best time is in the next five minutes.