AI engineer (LLM apps)
Learn to build, test and ship LLM-powered features and agents the way an AI engineering team does, using AI coding tools from week 1, in 20 weeks.
Coming soon · planned, not written yet
AI engineer is the fastest-growing job on LinkedIn’s 2026 listLinkedIn Jobs on the Rise 2026
- 20
- weeks planned
- ~40 h
- a week
- 5
- milestone projects
- 1
- project, grown every week
These are planned figures. The weeks are written after we check the tools against current job postings and official documentation, so the final route may differ.
Built for the AI era
How AI fits this roadmap
For this role AI is both the tool and the product. AI coding assistants now draft much of the Python, API and test code, so the job leans toward specifying, reviewing and testing what ships, and this roadmap has you work with an assistant every week and check its output with tests and evaluations. You also build the AI side itself: calling model APIs, retrieval over your own documents, tool use and agents, evaluation, and monitoring cost and quality in production.
The main project
DocPilot
An assistant that answers questions over a real document set, cites its sources, calls tools to take actions, and is tested, monitored and deployed as a web service.
What you will be able to do
- 01
Build a Python web API with FastAPI that calls a hosted model API, streams answers and returns structured, validated output.
- 02
Build a retrieval pipeline over a real document set, with chunking, embeddings and a vector store, and answers that cite the passages they used.
- 03
Write an evaluation set and run it on every change, so you can show whether a prompt, model or retrieval change made answers better or worse.
- 04
Build an agent that calls tools, including one MCP server you wrote, with limits on what it may do and a human check before risky actions.
- 05
Deploy the service with tracing, cost tracking and alerts, and write up one failure case with its cause and fix.
01The route
Five months, one month at a time.
Each month ends in a milestone project, so you always have something to show for the work.
Month 1 · Weeks 1–4
Python, APIs and working with an AI assistant
Typed Python, Git, tests and a FastAPI service. You set up an AI coding assistant, learn to give it clear instructions, read its diffs and test what it writes, then make your first model API calls with streaming and structured output.
MilestoneA tested API that sends a question to a model and returns validated, structured JSON, with a log of AI suggestions you accepted and rejected
Month 2 · Weeks 5–8
Prompting, context and retrieval
Prompt design, managing context, and retrieval over your own documents with embeddings and Postgres with pgvector. You use AI to draft the pipeline code, then check chunking and retrieval by hand.
MilestoneDocPilot answers questions over a document set and cites the passages behind each answer
Month 3 · Weeks 9–12
Evaluation and quality
Building an evaluation set, scoring answers with rules and with a model as judge, and checking where the judge is wrong. You try Ragas and compare it with your own checks, and learn when not to trust a model's answer.
MilestoneAn evaluation run on every change, with a report showing which changes helped and which hurt
Month 4 · Weeks 13–16
Tools, agents and MCP
Tool calling, agent loops, and the Model Context Protocol. You write an MCP server, limit what the agent may do, add human approval for risky actions, and use an AI assistant to attack your own agent with bad inputs.
MilestoneDocPilot calls tools through an MCP server you wrote, with tests for failures and refused actions
Month 5 · Weeks 17–20
Shipping, monitoring and capstone
Deployment, caching, cost limits, tracing with OpenTelemetry and Langfuse, and handling model errors and slow calls. Then a portfolio write-up, interview practice and applications.
MilestoneDocPilot live with tracing, cost tracking and alerts, plus a portfolio page and one failure write-up
03What is inside
Everything a week comes with.
Every week is built the same way, so once you have done one you know how all of them work.
- 01
One project, DocPilot
You grow one assistant from a first model call to a monitored service, so every tool lands in something real.
- 02
A day-by-day plan
Each week has a Learn part and a Do part, with a checklist so you know when you are done. Most weeks include a task to do with an AI assistant and a step to check its work.
- 03
A project in every milestone
Five milestones, each ending in something you can show and link to.
- 04
Reference topics
Short topic pages with check questions you can come back to.
- 05
A job track
Interview practice, portfolio and resume help, and an application plan.
- 06
Progress in your Study plan
Add weeks to your Study plan, give them dates, and finished days are counted for you.
04Is it for you
Who this is, and is not, for.
Better to know before you start than six weeks in.
It is for you if
- You can already write programs in some language, ideally Python, and want to build products on top of language models.
- You are a developer moving into AI work and want to be strong at working with AI coding tools.
- You can give about 40 hours a week for five months.
It is not for you if
- You have never written code. Do a beginner programming course first.
- You want to train models from scratch or do research. A machine learning roadmap fits that better.
- You want a job guarantee. Nobody can promise that, and this roadmap does not.
- You want to paste AI output into production without reading it. This roadmap trains you to check it.
?Questions
Before you start
- Is this roadmap ready?
- Not yet. This is the plan. The weeks are being written and checked against current job postings and official documentation. Leave your email and we will write to you when it opens.
- What will it cost?
- The price is not set. Roadmaps are bought once, with no subscription. We will announce it when this one opens.
- Will the plan change?
- It may. Before the weeks are written we check the tools against current job postings, so the lengths and months here are planned, not final.
- Do I need to pay for model APIs?
- Probably a small amount. You will call hosted model APIs, which charge by use. The roadmap will show how to keep the bill small and will name free options where they exist, once the weeks are written.
- Will it get me a job?
- It will give you skills and proof of work. It cannot promise a job on a fixed date, because hiring depends on the market, your location and your past work.
Not open yet
Leave your email and we will write once, when it opens.