Four structured tracks layered on top of Foundations. Each runs 4 weeks with concrete weekly outcomes, tools to learn, and a shippable artifact. Mix and match — most members do Fluency first, then one builder track.
T1
AI Fluency & Prompt Engineering
Use ChatGPT, Claude, and Gemini like a senior operator. Prompt patterns, evals, safe-use for confidential data.
ChatGPTClaudeGeminiPrompt librariesLMSYS Arena
- Week 1 — Model fluency: pick the right LLM per task; write 5 reusable prompt templates.
- Week 2 — Prompt patterns: role, few-shot, chain-of-thought, structured output (JSON).
- Week 3 — Evals & safe-use: build a 10-case eval set; redact PII; confidential-data checklist.
- Week 4 — Ship: a personal AI playbook (20+ prompts) used daily in your real work.
T2
No-Code / Low-Code AI Builders
Ship working AI products without writing backend code. Automations, agents, and full apps.
n8nMakeZapier AILovablev0Replit AgentBubble
- Week 1 — Automate one painful workflow end-to-end in n8n or Make.
- Week 2 — Build an AI agent that reads a doc, calls an API, and writes a result.
- Week 3 — Generate a full app UI with Lovable or v0; wire one real data source.
- Week 4 — Ship: a working AI tool you (or a paying user) actually use.
T3
Code & Data ML Foundations
The math-light path into real machine learning. Python notebooks, datasets, fine-tuning basics.
PythonGoogle ColabPandasscikit-learnHugging FaceKaggleFast.ai
- Week 1 — Python + Pandas in Colab: load a dataset, clean it, plot it.
- Week 2 — Train your first classifier with scikit-learn; read a confusion matrix.
- Week 3 — Run a Hugging Face model; fine-tune on a small custom dataset.
- Week 4 — Ship: a Kaggle submission or a Hugging Face Space demo of your model.
T4
AI App & Integration Engineering
Build production-grade AI features: APIs, RAG, agents, vector search, observability.
OpenAI APIGemini APILangChainLlamaIndexCrewAIPineconepgvector
- Week 1 — Call the OpenAI/Gemini API directly; stream responses; handle errors and cost.
- Week 2 — Build a RAG pipeline: embed docs, store in pgvector, retrieve, answer.
- Week 3 — Multi-step agent with LangGraph or CrewAI; add tools and guardrails.
- Week 4 — Ship: a deployed AI feature with logging, evals, and a cost dashboard.