Finguard ML · interactive course

Machine learning you can see, not just study.

A complete, hands-on course in machine learning and deep learning — from a single neuron to transformers, RAG, and production systems. Every idea is a figure you can adjust and watch respond. No walls of code; just understanding.

Start the course → Explore the curriculum
334
interactive lessons
41
units
Free
runs in your browser
Interactive figure · fit a line
Error (mean squared): —

A real figure from the course. Drag the sliders.

Reading about gradient descent is one thing. Watching the curve fall as you change the learning rate is another.

Finguard ML is built on a simple conviction: you understand an idea when you can play with it. So the entire syllabus is living, interactive figures — the kind you lose an afternoon to.

334
Interactive lessons
41
Units
9
Sections
100%
Hands-on, in-browser
Why it's different

Learning you can work with.

Most courses are a video and a wall of slides. We rebuilt the textbook into things that respond to you.

Figures, not lectures

Drag a decision boundary, slide a learning rate, watch a network train. Every concept ships with a hands-on figure that reacts instantly.

Built on real mathematics

Nothing is faked. Gradient descent actually descends, attention actually computes softmax, k-means actually converges — live in your browser.

A real path, end to end

Start at "what is a model" and finish at transformers, RAG, MLOps, and fairness — one coherent journey, each unit building on the last.

The curriculum

Nine sections, one path.

A guided arc from first principles to the frontier — 41 units and 334 lessons, each with worked examples and a knowledge check.

01
Machine Learning Foundations
What ML is, linear & logistic regression, working with data, generalization, and the metrics that matter.
6 units
02
Core ML Models
k-NN, decision trees, ensembles & random forests, SVMs, k-means clustering, and PCA — each one explorable.
6 units
03
Neural Networks & Deep Learning
From a single neuron to backpropagation, a training playground, CNNs with live convolution, and embeddings.
4 units
04
Sequence Models & Language
Tokenization, word embeddings, RNNs, the attention mechanism, transformers, LLMs, and retrieval-augmented generation.
4 units
05
Graphs & Advanced Topics
Knowledge graphs, graph neural networks, reinforcement learning in a gridworld, and generative models.
4 units
06
Real-world ML
Production systems, fairness & responsible AI, and recommendation systems — ML outside the notebook.
3 units
07
Applied ML & MLOps
Designing ML systems, data engineering, interpretability, A/B testing, time-series forecasting, and anomaly detection.
6 units
08
Applied Machine Learning: Projects
A complete ML roadmap, then five end-to-end projects on real datasets — tabular (Titanic), text (spam), images (digits), time series (sales), and recommenders (MovieLens).
6 units
09
Capstone Projects
Two integrative builds that tie the whole course together: credit-card fraud detection (imbalance, cost-sensitive thresholds, anomaly detection) and customer churn & retention (predict, explain, act, and A/B test).
2 units
Who it's for

For students, switchers, and researchers who want real intuition.

Students meeting ML for the first time Engineers moving into AI Researchers who want intuition fast Anyone tired of hand-wavy explanations Self-learners who like to tinker
Begin

Start learning in one click.

No account, no install. Your progress saves automatically in your browser. Pick any topic and start exploring.

Open the course → About Finguard