Learn AI Create the Future — with Pranay
Main site

All sections

The whole curriculum, grouped into 14 areas. Start at the top if you are new. If you came here for one thing, filter to its area or search — everything is on this page or one click from it.

Every lesson here is written by .

14 areas 117 sections 1,162 of 1,167 lessons published 10 new lessons every day

Showing Libraries and Frameworks — 4 sections, 40 lessons. Show all 14 areas

Libraries and Frameworks

The libraries you will actually type: scikit-learn, HuggingFace, TensorFlow and JAX.

4 sections 40 of 40 lessons published

scikit-learn Properly

The library most real tabular work is done in, used the way it was designed — pipelines first, so nothing leaks.

7 published

  1. The fit, predict, transform contract
  2. Pipelines and why leakage disappears
  3. ColumnTransformer for mixed data
  4. Writing your own transformer
  5. Encoders and scalers, and the unknown-category trap
  6. Choosing a cross-validation strategy
  7. Saving a scikit-learn model safely

The HuggingFace Stack

Beyond pipeline(): the model classes, tokeniser details, datasets and trainers you need to build something real.

13 published

  1. Which AutoModel class to use
  2. Fast tokenizers, offsets and word_ids
  3. Padding, truncation and attention masks
  4. Controlling generate()
  5. Loading models bigger than your GPU
  6. The Hub cache and pushing your own model
  7. The Datasets library
  8. map, batched and the cache
  9. Data collators
  10. The Trainer API
  11. Accelerate
  12. Loading a model in 4-bit
  13. QLoRA: fine-tuning a large model on one GPU

TensorFlow and Keras

For teams already on TensorFlow, and for anyone who inherits a Keras codebase and needs it to make sense.

10 published

  1. Sequential and functional models
  2. compile, fit and evaluate
  3. Callbacks
  4. Custom layers, losses and metrics
  5. Overriding train_step
  6. GradientTape
  7. tf.function, graphs and retracing
  8. tf.data input pipelines
  9. Saving and loading Keras models
  10. Stopping TensorFlow taking the whole GPU

JAX and Flax

A different way to think about numerical code: pure functions in, transformed functions out, then compiled to run fast.

10 published

  1. JAX arrays and immutability
  2. jit and tracing
  3. grad and value_and_grad
  4. vmap: write the code for one example
  5. Pytrees
  6. Random numbers and PRNG keys
  7. Control flow: scan, cond and while_loop
  8. Building models with Flax
  9. Optax
  10. A full training loop in JAX

Other ways to use this site

Learning paths

An ordered route through the lessons for one job: ML developer, AI engineer, computer vision engineer.

Projects

Build real things with full code, a dataset and a deployment guide.

AI glossary

Every term you keep seeing, defined in one plain sentence first.

Error database

Paste the error you got. Find out what it means and how to fix it.