From @pranjulya.si. One free resource for every step, plus a 6-month plan.
<aside>
⏱️
The plan assumes about 10 hours a week alongside a job. Backend dev? Skip Month 1 and start at step 4.
</aside>
Month 1: Foundations
- Python: The official Python tutorial
- APIs: FastAPI tutorial (build a small REST API)
- SQL: SQLBolt (interactive lessons)
Month 2: How AI works
- ML basics: Google Machine Learning Crash Course
- Neural nets: 3Blue1Brown: Neural networks
- Tokens + context: play with the OpenAI tokenizer, then watch Andrej Karpathy's "Let's build the GPT Tokenizer" on YouTube
- Embeddings: The Illustrated Word2vec (Jay Alammar)
Months 3–5: Build (the most important part)
- Prompting: Prompt engineering overview (Claude docs)
- Structured output: Structured outputs guide (OpenAI docs)
- RAG: Introducing Contextual Retrieval (Anthropic)
- Tools + MCP: Model Context Protocol docs
- Agents: Building effective agents (Anthropic)
Month 5: Production
- Evals: Your AI product needs evals (Hamel Husain)
- Tracing: Langfuse docs (open source LLM tracing)