Skip to content

Study Plans

Concrete week-by-week schedules by persona. For the full ordered list of courses, see Learn.

Choose your path

Persona Background Time/week Duration Jump to
Beginner Software engineer, little ML 8–10 hrs ~20 weeks Beginner plan ↓
Intermediate Know ML/Python, new to LLMs & agents 6–8 hrs ~12 weeks Intermediate plan ↓
Agent engineer Shipping or designing agent systems 5–7 hrs ~8 weeks Agent engineer plan ↓

Beginner path (~20 weeks)

Goal: Understand transformers and LLMs, build a RAG app, touch agents, learn production basics.

Week Focus Courses Milestone
1–2 NLP, attention, transformers intro 01 GenAI Foundations Sketch a transformer block
3 First API app 02 AI Essentials Working chat script
4–6 Neural nets from scratch 03 Neural Networks Train a small classifier
7–8 Attention deep dive 04 Transformers Compute attention on toy tokens
9–10 LLM lifecycle 05 LLMs Explain pretrain → SFT → RLHF
11–12 RAG 06 RAG Local doc Q&A
13 Prompts 11 Prompt Engineering A/B two system prompts
14–15 Agents 07 AI Agents ReAct loop with memory
16 Vector DBs 10 Vector Databases Tune retrieval on your corpus
17 LLMOps 12 LLMOps Logging + caching on RAG app
18 Evals 13 LLM Evaluation 10 golden cases in CI
19 Safety 14 AI Safety Red-team one prompt
20 Capstone prep 16 Capstones Project proposal doc

Optional any week: Deep Dives

Intermediate path (~12 weeks)

Skip classical ML depth if you already know it; focus on LLMs, RAG, agents, production.

Weeks Focus Courses
1–2 LLM essentials 02, 05 (+ Attention deep dive)
3–5 Build systems 06 RAG, 11 Prompts, 10 Vector DBs, 07 Agents + Agent Engineering
6–8 Agents in depth 08 Harness, 09 Multi-Agent, Agent Engineering + Modern AI 2026
9–12 Production + advanced 12 LLMOps, 13 Evals, 14 Safety, 15 Fine-Tuning or 16 Capstones

Milestone (week 8): One agent with harness, traces, and 5 golden trajectories in CI.

Agent engineer path (~8 weeks)

Goal: Production-grade agents — harness, orchestration, observability, evals — fast.

Week Focus Resources
1 Loop + harness Agent Loop, 07 Agents, 08 Harness
2 Tools, MCP, memory Tools & MCP, Memory, Context Engineering
3 Orchestration Orchestration, 09 Multi-Agent
4 Observability Observability
5 Evals Agent Evals, 13 LLM Evaluation
6 2026 tooling Modern AI 2026
7 Production 12 LLMOps, safety lessons in 08 + 14
8 Integration Harness + traces + evals + runbook (see Build These)

Study habits

  1. One lesson → one artifact — notebook, script, or eval case per lesson
  2. Teach it — explain attention or ReAct aloud in 5 minutes
  3. Use the glossary — Glossary when terms blur together
Stuck on… Go to
Math Deep Dives
Agent loops forever Harness termination
Bad RAG answers 06 RAG · Evaluation

Start: Pick a persona above and block Week 1 on your calendar today.