Machine Learning
Topic 1 · Getting started · ⏱ about 1 h

Your Roadmap: Zero to ML Intern

What employers expect from an ML intern or junior, how long it takes, and a week-by-week plan through this course.

Show:
In this topic
Beginner

Where you are going

This course is designed so that someone with no programming and no university maths can reach the level expected of a machine learning intern or junior ML engineer. That is a realistic goal, but it is not a quick one: plan for roughly 300–400 hours of focused work, which is 8–10 months at 10 hours a week, or 4–5 months at 20.

What do hiring managers actually look for in an ML intern or junior? Across job ads and interview loops, the same things come up:

  • Comfortable Python, including NumPy and pandas, and enough general programming to pass an easy coding interview.
  • The everyday tools: the terminal, git and GitHub, Jupyter, virtual environments, and writing simple tests.
  • SQL to get your own data.
  • Solid fundamentals: train/validation/test splits, leakage, overfitting, the right metric, baselines, and the main model families and when to use them.
  • One or two end-to-end projects on real data, on GitHub, with a README that explains the result.
  • Judgement and communication: explaining a model's limits, its errors and its risks to people who are not engineers.

Certificates matter much less than a portfolio and a good interview. Everything in this roadmap is aimed at those.

Beginner

The seven phases

  1. Programming foundations: Python from zero, NumPy, pandas, and the developer toolkit (terminal, git, environments, tests). Do not skip this. Every later exercise assumes it.
  2. Maths and statistics: notation, vectors, matrices, derivatives, probability, likelihood, eigenvectors, optimisation, and statistics for decisions (confidence intervals, tests, A/B testing).
  3. Core machine learning: data, SQL, cleaning, features, linear and logistic regression, gradient descent, overfitting, regularisation and evaluation, with the first two game milestones.
  4. Algorithms: k-NN, trees, ensembles, SVMs, clustering, PCA, time series, recommenders and anomaly detection.
  5. Deep learning: neural networks, training, CNNs/RNNs/transformers, and PyTorch notebooks on a real GPU.
  6. Production and LLMs: reinforcement learning, the project workflow, deployment (APIs, Docker, tracking, monitoring), LLMs and RAG.
  7. Job ready: the capstone, two portfolio projects, Python interview practice, and the interview-day milestone.

Each phase ends with a checkpoint: a short mixed quiz. If you score below the pass mark it tells you exactly which topics to revisit before moving on. Use the planner below to turn the phases into dates.

Try it

Try it: plan your weeks

Choose how many hours a week you can study and when you start. The plan uses the time estimate shown at the top of every topic, and fills in as you complete topics and checkpoints.

Interactive: this part runs in your browser.

Beginner

How to study so it sticks

  • Type the code yourself. Reading a solution feels like understanding; typing and debugging it is understanding. Use hints before solutions.
  • Little and often beats cramming. Five sessions of two hours beat one ten-hour day. Return to quizzes and flashcards a few days later (spaced repetition).
  • Explain it out loud. After each topic, explain the key idea in two minutes as if to a friend. If you get stuck, that is the part to re-read.
  • Build things early. From phase 3, start a small project of your own alongside the course, even if it is rough.
  • Check facts against primary sources. Each topic ends with further reading from trusted sources (official documentation, well-known textbooks and courses): when something here surprises you, check it there.
  • Join a community. A study group, a local meetup, or Kaggle discussions. Explaining and asking questions accelerates learning and builds the network that finds jobs.
Try it

Try it: am I intern-ready?

Tick each skill only if you could do it today without looking anything up beyond documentation. Your answers are saved in this browser. Untick-ed skills link to the topic that teaches them.

Interactive: this part runs in your browser.

Try it

Back up your progress

Your progress (topics done, checkpoint scores, game state, code you wrote) is stored in this browser only. To move it to another browser or computer, copy the backup text here and paste it into the same box there.

Interactive: this part runs in your browser.

Check yourself

Quiz

1. Which matters most when applying for an ML internship?

2. You have 10 hours a week. Roughly how long will this roadmap take?

3. What should you do when something in this course seems wrong or surprising?

4. Which study habit helps most with retention?

Go deeper

📚 Further reading

Trusted, mostly free sources: official documentation, well-known textbooks and courses. When you want to confirm or go deeper on anything here, start with these.

Ready