Machine learning engineer
Learn to train, evaluate and ship machine learning models, and to build with today's AI models, in 24 weeks, using AI coding tools from week 1.
Coming soon · planned, not written yet
AI and ML roles lead LinkedIn’s 2026 rising-jobs listLinkedIn Jobs on the Rise 2026
- 24
- weeks planned
- ~30 h
- a week
- 6
- 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
AI coding assistants now write much of the first draft of data and model code, so the skill that matters is reviewing it, testing it and knowing when it is wrong. From week 1 you work with an assistant and learn to check its output against tests and the maths. Later weeks cover the AI that is the job itself: using pretrained and large language models, fine-tuning them, and measuring whether they actually work.
The main project
Signal Desk
A model-backed service that classifies and ranks incoming support tickets, which you grow from a notebook into a tracked, tested, deployed API with a language-model feature added and measured.
What you will be able to do
- 01
Clean and analyse real tabular data with Python, NumPy and pandas, and explain the results with the basic maths behind them.
- 02
Train and compare classical models with scikit-learn, and choose metrics that fit the problem rather than the default.
- 03
Build, train and debug a neural network in PyTorch, and fine-tune a pretrained model from Hugging Face Transformers.
- 04
Review AI-written code with tests, catch wrong output, and write a short evaluation set that shows whether a language-model feature works.
- 05
Track experiments with MLflow, serve a model through a FastAPI service in a Docker container, and write up the result for a portfolio.
01The route
Six 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, data and the maths you need
Python for data work with NumPy and pandas, plus the vectors, probability and calculus ideas behind models. You use an AI assistant daily and learn to test what it writes.
MilestoneA cleaned dataset and a notebook that explains it, with tests on your data code
Month 2 · Weeks 5–8
Classical machine learning
Regression, classification, trees and ensembles with scikit-learn. Splits, leakage, cross-validation and metrics that match the problem. You ask an AI assistant for model code, then check it for leakage and wrong metrics.
MilestoneA baseline ticket classifier with an honest evaluation report
Month 3 · Weeks 9–12
Neural networks with PyTorch
Tensors, training loops, overfitting and debugging in PyTorch. You check AI-generated training code against what you know about gradients and loss.
MilestoneA neural network trained from your own loop and compared against the baseline
Month 4 · Weeks 13–16
Language models and fine-tuning
How transformers work, using pretrained models from Hugging Face Transformers, fine-tuning one for your task, and calling a hosted language model for a second feature.
MilestoneA fine-tuned ticket model plus a language-model summary feature, both with an evaluation set
Month 5 · Weeks 17–20
Tracking, serving and shipping
Experiment tracking with MLflow, packaging a model behind a FastAPI service, Docker, automated tests and a basic check for drift after release.
MilestoneSignal Desk running in a container, with tracked experiments and tests on every change
Month 6 · Weeks 21–24
Capstone and job search
A final improvement cycle on Signal Desk, a clear write-up of choices and limits, interview practice for ML roles, and applications.
MilestoneA portfolio write-up, a public repository and a working job-search pipeline
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, Signal Desk
You grow one model-backed service from notebook to deployed API, so every method 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.
- 03
A project in every milestone
Six milestones, each ending in something you can show and link to.
- 04
Reference topics
Short topic pages on the maths, metrics and tools, 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 write small programs in Python or another language and want to move into machine learning.
- You are a developer, analyst or data worker who wants to build models, not only call them.
- You can give about 30 hours a week for six months. This is a part-time pace.
It is not for you if
- You have never written code. Do a beginner programming course first.
- You want a research career or a PhD-level theory course. This roadmap teaches the maths you need to build and judge models.
- You want a job guarantee. Nobody can promise that, and this roadmap does not.
?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 a GPU or a paid cloud account?
- The plan is to keep the work runnable on a normal laptop, using small models and free or low-cost hosted notebooks for the heavier training weeks. The roadmap will say before each week what you need.
- 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.