Understand machine learning by building it

graded problems
168
graded problems
topics
43
topics
to start
$0
to start

Based on analysis of:

GoogleMetaAmazonAppleMicrosoftTeslaSpaceXStripeNVIDIAOpenAIAnthropicDatabricks
GoogleMetaAmazonAppleMicrosoftTeslaSpaceXStripeNVIDIAOpenAIAnthropicDatabricks

Built for people who learn by shipping code

Less theory theatre. More working implementations you can test, debug, and actually remember.

normalize.py
01

Practice problems

A real Python editor, graded against fixed test cases with known-correct answers.

Fraud Detection

Precision0%
Recall0%
held-out metrics passed
02

Industry projects

Longer builds modeled on real ML work, graded on deterministic checks and held-out metrics.

AI
Here's where you stand after this week's submissions:
Gradient Descent92%
Backpropagation78%
Regularization50%
03

AI Mentor

Chats with you about your own submitted code and shows exactly where you stand.

Attention Is All You Need

Attention(Q,K,V) = softmax(QKT/√d)V

def attention(Q, K, V):

s = Q @ K.T / d**0.5

return softmax(s) @ V

04

Built from real papers

Advanced problems replicate real techniques, causal attention masking, LSTM gates, gradient boosting, not simplified toy versions.

A structured path through core ML

From NumPy foundations to deep learning. Open any concept and start coding.

NumPy Arrays

Foundations

Gradient Descent

Classical ML

Backpropagation

Deep Learning

Regularization

Deep Learning

Adam Optimizer

Deep Learning

Model Evaluation

Classical ML

Frequently asked questions

What is GRADuateML?
GRADuateML (also searched as Graduate ML or graduate ml) is an interactive machine learning education platform at graduateml.com. You understand ML by writing real Python code in the browser, getting instant test feedback, and building mastery across concepts from NumPy to neural networks.
Is Graduate ML the same as GRADuateML?
Yes. GRADuateML is the official brand name. People often search for Graduate ML, graduate ml, or graduateml - they all refer to the same platform at graduateml.com.
Where can I find machine learning practice problems?
GRADuateML offers 168 free machine learning practice problems and Python coding questions across 43 concepts, from gradient descent to neural networks. Browse the curriculum at graduateml.com/learn and start solving problems in your browser without installing anything.
Is GRADuateML free?
Yes. You can sign up free and start solving coding problems immediately. Pro unlocks the AI Mentor, unlimited daily submissions, and advanced curriculum sections.
Do I need to install anything?
No. GRADuateML runs entirely in your browser with a full code editor. You write Python, run tests, and get feedback without local setup.
What ML topics does GRADuateML cover?
The curriculum spans foundations (NumPy, linear algebra), classical ML (gradient descent, logistic regression, model evaluation), and deep learning (backpropagation, regularization, optimization). Each concept includes learning objectives and hands-on coding problems.
How do AI hints work?
When you are stuck on a Pro problem, you can request progressive hints: a conceptual nudge, a more specific hint, or a full explanation. The AI analyzes your submitted code and diagnoses your specific mistake.

Start with a free account

Jump into practice problems, track mastery, and build interview-ready ML intuition.