Choose your path: 3, 6 or 12-month plans

Why this matters #

Most self-taught learning plans fail the same way: an intense first week, a busy second week, and then nothing. The fix isn’t more motivation. It’s a plan sized to the hours you can actually protect, with a clear order and a rule for what to cut when life gets in the way.

This article helps you check you’re ready, choose one of three plans, and set up a weekly routine that survives a bad week.

Before you start #

The wiki assumes you can already build a small web app. Check yourself against this list:

• You can write TypeScript or JavaScript comfortably, including async and await.
• You’ve built something that calls an HTTP API and handles errors.
• You can run a project locally, use git, and deploy somewhere, even a free tier.
• You’ve used a framework such as Next.js or React, or can pick one up as you go.

If two or more of those are shaky, spend a few weeks on web fundamentals first. If you’ve never called an LLM API, read the opening stages of the AI Full-Stack Engineer Wiki (opens in a new tab) alongside stages 1 to 3 here.

You don’t need machine learning experience, a maths background or a paid AI subscription. You will need an Anthropic API key for the code examples and projects. Most examples cost a few cents to run, but check current prices: this site uses Claude Opus 5.5 at $4 input / $20 output per million tokens by default.

The three plans #

All three plans cover the same nine stages in the same order. They differ in pace and in how many of the six projects they include.

3-month plan 12–15 h a week · 3 projects

6-month plan 7–9 h a week · 5 projects

12-month plan 4–5 h a week · 6 projects

Week 113263952

Beginner stages Intermediate stages Advanced stages Expert stages

Show text descriptionHide text description

Three horizontal bars on a shared scale of 52 weeks.

• 3-month plan (13 weeks, 12–15 hours a week): weeks 1–2: Start Here and Product Foundations; weeks 3–5: LLM Capabilities for Product Decisions and Designing AI UX plus project 1; weeks 6–8: Rapid Prototyping and Evals as Product Metrics plus project 2; weeks 9–11: Shipping & Iterating and Trust, Safety & Economics; weeks 12–13: Projects & Career plus project 6 .
• 6-month plan (26 weeks, 7–9 hours a week): weeks 1–3: Start Here and Product Foundations; weeks 4–8: LLM Capabilities for Product Decisions and Designing AI UX plus project 1; weeks 9–12: Rapid Prototyping plus project 2; weeks 13–16: Evals as Product Metrics plus project 3; weeks 17–20: Shipping & Iterating plus project 5; weeks 21–23: Trust, Safety & Economics; weeks 24–26: Projects & Career plus project 6 .
• 12-month plan (52 weeks, 4–5 hours a week): weeks 1–5: Start Here and Product Foundations; weeks 6–14: LLM Capabilities for Product Decisions and Designing AI UX plus project 1; weeks 15–20: Rapid Prototyping plus project 2; weeks 21–28: Evals as Product Metrics plus project 3; weeks 29–36: Shipping & Iterating plus project 4 and 5; weeks 37–42: Trust, Safety & Economics; weeks 43–52: Projects & Career plus project 6 .

Hours are estimates for an experienced developer. Take longer if you're new to TypeScript or product work: the order matters more than the pace.

How to choose #

Answer three questions honestly:

1. How many hours can you protect every week, including bad weeks? Not your best week: your typical one. Pick the plan whose lower bound fits.
2. Do you have a deadline? A job search, a launch or a review cycle can justify the 3-month plan, if you have the hours.
3. How much of this do you already do? If you already ship AI features at work, the faster plans are realistic because the projects overlap with your job.

When in doubt, choose the slower plan. Finishing the 12-month plan beats abandoning the 3-month one in week five, and you can always speed up. Switching to a faster plan mid-way is easy because the order never changes.

A week that works #

Here’s a routine for the 6-month plan. Scale the blocks up or down for the other plans.

A few things make this routine stick:

• Put the blocks in your calendar at the same times every week. A fixed slot beats “when I have time.”
• Do the quiz even when you feel confident. It’s the fastest way to find what didn’t stick.
• Keep the reflection notes. They become raw material for your case studies in Projects & Career.
• Build in public if it helps you. A short weekly post about what you learned creates gentle accountability, and a record employers can read.

When you fall behind #

You will have weeks where nothing happens. Plan for them with three rules:

1. Cut projects, not stages. Each stage builds on the previous one, but the plans already skip projects. Dropping one more is fine. Skipping a stage isn’t.
2. Never skip evals. Evals as Product Metrics is the stage people most want to skip and the one that most separates AI product engineers from people who call APIs.
3. Restart small. After a missed week, don’t try to catch up in one sitting. Do one article and its quiz, then return to the normal routine.

If you fall behind by more than three weeks, move to the next slower plan. It’s not failure: the stages and order are the same, so you keep everything you’ve done.

Common mistakes #

Try it #

Exercise

About 20 min

Commit to a plan and book the time

1. Look at the last four weeks in your calendar and count the hours you could genuinely have protected each week.
2. Choose the plan whose lower bound fits your typical week, not your best one.
3. Book the read, build and reflect blocks as recurring calendar events for the next four weeks.
4. Write a one-paragraph “why” note: what you want to be able to do in six months, and why it matters to you. Save it somewhere you’ll see it.

Show a hint

If your count varies a lot week to week, use the second-lowest week, not the average. Plans fail in the hard weeks, so size them for those.

Show a worked solution

Example: a developer finds they protected 9, 4, 8 and 7 hours over the last four weeks. The second-lowest week is 7 hours, which fits the 6-month plan’s 7–9 hours a week. They book Tuesday and Thursday evenings for reading, Saturday morning for building, and Sunday evening for a 15-minute reflection. Their “why” note: “In six months, I want to own an AI feature at work end to end, with evals and metrics I can show in my next performance review.”

Check your understanding #

Key takeaways #