A written version of the interactive roadmap above — every station, what you'll learn, and a small thing to build — laid out for reading, reference and search.
Foundations Start here
F1. Math for ML
Beginner · 75 min
Machine learning is applied math. You do not need a PhD, but you do need the core: linear algebra (vectors, matrices), calculus (gradients — how models learn), and probability (uncertainty and likelihood). These three show up in every model you will ever train.
Skills: Vectors & matrices · Gradients & derivatives · Probability & distributions · Dot products & norms
Build it: Compute a gradient by hand for a tiny function, then explain what "descending the gradient" means for a model chasing lower loss.
✓ Checkpoint: Explain what a gradient tells you geometrically, and why that is enough to train anything at all.
F2. Python, NumPy & pandas
Beginner · 60 min
Python is the language of ML, and NumPy/pandas are its workhorses. Learn vectorized array math (loops are the enemy of speed), dataframes for tabular data, and the scikit-learn API shape. Fluency here is what lets you go from idea to experiment in minutes.
Skills: NumPy arrays & vectorization · pandas dataframes · scikit-learn API · Plotting & EDA
Build it: Vectorize a slow Python loop with NumPy and measure the speedup. Why is the array version often 100x faster?
✓ Checkpoint: Explain why vectorised NumPy beats a Python loop by so much that the difference changes what is feasible.
F3. Data Prep & Feature Engineering
Intermediate · 65 min
Models are only as good as their features. Learn cleaning, handling missing values, encoding categoricals, scaling, and crafting features that expose signal. In classical ML, thoughtful feature engineering beats a fancier model more often than not.
Skills: Cleaning & missing values · Encoding categoricals · Scaling & normalization · Feature creation & selection
Build it: Turn a raw table with dates, categories, and nulls into model-ready features. Which engineered feature would most help a churn model?
✓ Checkpoint: Explain why fitting your scaler on the full dataset before splitting leaks the test set — and how the leak shows up as an optimistic score.
F4. Supervised Learning
Intermediate · 70 min
The dominant paradigm: learn a mapping from inputs to labeled outputs. Learn the split between classification and regression, training vs test data, the bias-variance tradeoff, and overfitting. This mental model underpins nearly everything else in the roadmap.
Skills: Classification vs regression · Train / validation / test · Bias-variance tradeoff · Overfitting & regularization
Build it: A model scores 99% on training and 70% on test. Name the problem, and three ways to close the gap.
✓ Checkpoint: State the bias-variance position of a model that scores 99% on train and 70% on test, and what you would change.
F5. Unsupervised Learning
Intermediate · 55 min
When there are no labels, structure must be found in the data itself. Learn clustering (k-means), dimensionality reduction (PCA), and embeddings — the tools for exploration, compression, and finding hidden groups. This is also where representation learning begins.
Skills: Clustering (k-means) · Dimensionality reduction (PCA) · Embeddings · Anomaly detection
Build it: You have customers but no segments. Cluster them, then reduce to 2D to visualize. How do you decide how many clusters are real?
✓ Checkpoint: Explain how you would judge whether k-means found real structure or just partitioned noise.
F6. Model Evaluation
Intermediate · 60 min
A single accuracy number lies. Learn the metrics that matter — precision, recall, F1, ROC-AUC for classification; MAE/RMSE for regression — cross-validation, and why the right metric depends on the cost of each kind of error. Evaluation is where honesty lives.
Skills: Precision / recall / F1 · ROC-AUC & confusion matrix · Cross-validation · Choosing the right metric
Build it: A fraud model is 99.9% accurate but useless. Explain why, and which metric you should have optimized instead.
✓ Checkpoint: Explain why accuracy is the wrong metric on a 1%-positive dataset, and which metric you would report instead.
Core ML The craft
T1. Linear & Logistic Regression
Intermediate · 60 min
Start where ML started. Linear regression fits a line; logistic regression bends it into a probability for classification. Learn the loss functions, gradient descent, and regularization. These simple, interpretable models are still the right first answer to a huge share of problems.
Skills: Linear regression · Logistic regression · Loss & gradient descent · L1 / L2 regularization
Build it: Fit logistic regression to a binary problem, then read the coefficients. What does a large positive weight tell you about a feature?
✓ Checkpoint: Explain what logistic regression actually outputs, and why calling it a probability requires calibration.
T2. Trees & Ensembles
Advanced · 65 min
Decision trees split data into rules; ensembles combine many weak trees into a strong model. Learn random forests (bagging) and gradient boosting (XGBoost/LightGBM). On tabular data, boosted trees remain the models to beat — often outperforming deep nets.
Skills: Decision trees · Random forests (bagging) · Gradient boosting (XGBoost) · Feature importance
Build it: A boosted-tree model wins on your tabular data. Explain why trees often beat neural nets here — and what feature importance reveals.
✓ Checkpoint: Explain the difference between bagging and boosting in terms of what each one is trying to reduce.
T3. Neural Networks & Backprop
Advanced · 75 min
Neural nets stack simple units into powerful function approximators. Learn neurons, activation functions, forward passes, and backpropagation — the algorithm that assigns credit and lets a network learn. Understand backprop once and every deep model stops being magic.
Skills: Neurons & layers · Activation functions · Forward pass · Backpropagation
Build it: Trace one training step through a 2-layer net: forward to a loss, then backprop the gradient. Where does the chain rule appear?
✓ Checkpoint: Explain what backpropagation is doing with the chain rule, and why the forward pass has to be remembered.
T4. Deep Learning
Advanced · 75 min
Depth unlocks representation learning — features are learned, not hand-crafted. Learn training at scale: optimizers (Adam), batching, learning-rate schedules, regularization (dropout), and the frameworks (PyTorch) that make it practical. This is the gateway to modern AI.
Skills: Optimizers (SGD/Adam) · Batching & epochs · Dropout & batch norm · PyTorch basics
Build it: Your deep net will not converge. Walk through learning rate, initialization, and normalization as the usual suspects — which do you check first?
✓ Checkpoint: Explain what dropout does at training time and what it must NOT do at inference.
T5. CNNs & Computer Vision
Advanced · 65 min
Convolutional networks brought the deep-learning revolution to images. Learn convolutions, pooling, feature hierarchies, and landmark architectures (ResNet). Even in a transformer world, CNNs remain the efficient default for many vision tasks — and the intuitions transfer.
Skills: Convolutions & filters · Pooling & feature maps · Architectures (ResNet) · Transfer learning
Build it: Fine-tune a pretrained ResNet on a small image dataset. Why does transfer learning beat training from scratch with little data?
✓ Checkpoint: Explain what a convolution buys over a fully-connected layer for images — the answer is about structure, not size.
T6. Transformers & NLP
Advanced · 75 min
Transformers replaced recurrence with attention and now dominate language — and beyond. Learn tokenization, embeddings, self-attention, and the encoder/decoder split that underpins BERT and GPT. This is the architecture behind every modern large language model.
Skills: Tokenization & embeddings · Self-attention · Encoder vs decoder · Pretraining & fine-tuning
Build it: Explain self-attention as "each token looks at every other token." Why does that beat a fixed-window model for long-range meaning?
✓ Checkpoint: Explain why self-attention is quadratic in sequence length, and what that costs on a long document.
Production Ship & operate
P1. Experiment Tracking
Advanced · 50 min
ML without tracking is chaos — dozens of runs and no memory of what worked. Learn experiment tracking (metrics, params, artifacts), data and model versioning, and reproducibility. The goal: any result can be reproduced, and the best model is never a mystery.
Skills: Experiment tracking · Data & model versioning · Reproducibility · Config management
Build it: You beat your baseline last week but cannot reproduce it. What should you have logged — data version, seed, params — to get it back?
✓ Checkpoint: Explain why an experiment you cannot reproduce is not a result yet.
P2. MLOps & Deployment
Advanced · 65 min
A model in a notebook helps no one. Learn to serve models — batch vs real-time inference, packaging, APIs, and CI/CD for models. MLOps brings software discipline to ML so models ship reliably and repeatedly, not once by hand.
Skills: Batch vs real-time serving · Model packaging & APIs · CI/CD for models · Rollout & rollback
Build it: Turn a trained model into a low-latency API. Where do you validate inputs, and how do you roll back a bad model version fast?
✓ Checkpoint: Explain what has to be true of a model before it can be served, beyond the metric it was selected on.
P3. Feature Stores
Advanced · 50 min
Features used in training must match features at serving time, or the model silently breaks. A feature store centralizes feature definitions, serves them online and offline, and prevents training-serving skew — a subtle bug that quietly wrecks production models.
Skills: Online vs offline features · Training-serving skew · Feature reuse · Point-in-time correctness
Build it: A model is great offline but poor live. Show how a feature computed differently in training vs serving causes it — and how a store fixes it.
✓ Checkpoint: Explain training-serving skew with a concrete example, and how a feature store prevents it.
P4. Monitoring & Drift
Advanced · 60 min
Models decay because the world changes. Learn to monitor prediction quality, detect data drift and concept drift, and trigger retraining. Unlike a service that stays broken until fixed, a model degrades quietly — monitoring is what catches the slow rot.
Skills: Data vs concept drift · Prediction monitoring · Retraining triggers · Shadow & canary models
Build it: Accuracy is fine but the input distribution shifted. Which drift metric fires first, and when does that justify a retrain?
✓ Checkpoint: Explain the difference between data drift and concept drift, and why the second is harder to detect.
P5. Distributed Training
Advanced · 60 min
Big models and big data outgrow one GPU. Learn data parallelism vs model parallelism, gradient synchronization, mixed precision, and the memory/throughput tradeoffs. This is how modern models are trained across many accelerators without falling over.
Skills: Data vs model parallelism · Gradient synchronization · Mixed precision · GPU memory & throughput
Build it: A model no longer fits on one GPU. Choose data vs model parallelism and explain the communication cost each strategy adds.
✓ Checkpoint: Explain the difference between data and model parallelism in terms of what gets split across devices.
P6. Responsible AI
Advanced · 55 min
Powerful models carry real risk. Learn fairness and bias auditing, interpretability (SHAP, feature attribution), privacy, and the human oversight that keeps automated decisions accountable. Responsible AI is not a compliance afterthought — it is part of the engineering.
Skills: Fairness & bias auditing · Interpretability (SHAP) · Privacy & data ethics · Human oversight
Build it: A loan model rejects one group more often. How do you measure whether it is biased, and what levers reduce the disparity?
✓ Checkpoint: Explain what SHAP tells you and what it does not — an explanation is not a justification.