Open-source

Become an AI engineer,
one milestone at a time.

A structured, hands-on path from fundamentals to production AI systems — prompting, RAG, agents, evals, and fine-tuning. Then a built-in AI tutor drills you on every milestone with flashcards, mock interviews, and spoken-explanation grading.

Free & open-sourceNo signupLive AI tutor

Why this roadmap

Learn the path. Get drilled by an AI tutor.

It's a curriculum and a coach — the content teaches, the tutor makes it stick.

A structured path, not a link dump

108 prerequisite-ordered lessons, fundamentals → production — prompting, RAG, agents, evals, and fine-tuning, sequenced so every milestone builds on the last.

A tutor, not just docs

A live AI tutor generates flashcards, runs mock interviews, and grades your spoken explanations — so you practice, not just read.

Always fresh

Every lesson is RAG-grounded in primary sources and kept current by an autonomous knowledge-graph pipeline that runs on a schedule.

Featured hubs

Pick a track. Start building.

Deep-dive hubs — hands-on guides, audio walkthroughs, and flashcards.

How It Works

How this very site is built and grounded — content pipeline, local-first embeddings, egress tiers, and deploy. Audio tab: the narrated platform guide.

Explore

RAG & Retrieval

Embeddings, vector search, and grounded answers — with audio guides and podcasts.

Explore

Evals & Testing

Ship safely: evals, red-teaming, and guardrails that prove it works.

Explore

LangGraph

Stateful agent graphs with checkpointed memory and human-in-the-loop.

Explore

LlamaAgents & Workflows

Event-driven agent workflows: steps + events + Context, served, made durable, and deployed.

Explore

Agent Frameworks Compared

LangGraph vs LlamaIndex Workflows vs OpenAI Agents SDK — the same parallel-agent system built three times, measured, with the footgun each framework hides.

Explore

Human-in-the-Loop

How eight frameworks implement the pause — interrupt(), InputRequiredEvent, human_input, needs_approval — compared primitive by primitive, with each one's persistence and resume story.

Explore

Agent Autonomy

The autonomy spectrum — agent loops, human-in-the-loop, durable execution, and guardrails.

Explore

LangSmith

Tracing and observability for production LLM apps.

Explore

Apply vs. Learn

When to apply, how hard, for how long — and when to step back and upskill instead. A research-grounded job-search cadence framework.

Explore

Daily Routine

The day-to-day playbook: time-boxed apply/outreach/learn blocks, a printable checklist, a sample day, and the weekly review that picks next week's mode.

Explore

Memorize

Drill the core concepts with an AI tutor that generates spaced-repetition flashcards on the fly.

Explore

Mock Interviews

A conversational AI interviewer that questions you on the curriculum, then grades your answers.

Explore

Voice Tutor

Explain a concept out loud; a Feynman-style evaluator scores your understanding against the reference.

Explore