Where to Go from Here

You Have a Foundation

Eleven tracks, 59 lessons, and forty quizzes have given you a working map of the mathematics that repeatedly appears in machine learning. It is enough to make the next source readable, not a claim that every paper will now be easy.

Immediate Next Steps

Machine Learning Foundations Quest — supervised and unsupervised methods, evaluation, and classical algorithms.
Deep Learning Foundations Quest — architectures, optimization, regularization, and training systems beyond the toy examples here.
PyTorch Quest — tensors, modules, data pipelines, custom training loops, and distributed execution.

Specializations

Transformer Quest — attention, residual streams, normalization, feed-forward blocks, training, and decoding.
Vector Quest — embeddings, retrieval, semantic search, and RAG.
Eval Quest — metrics, test design, calibration, human judgment, and failure analysis.
Beyond the Transformer / Beyond Attention — state-space models, mixtures of experts, and alternative sequence mechanisms.

Put the Map Under Load

Choose a small project, trace a paper's assumptions, or implement a tiny autograd engine. When a formula appears, write the shapes, domain, and role before running it. Practice turns the map into judgment.

Keep concepts and formal checks together. Build something small enough to inspect, then let the next real failure choose what you study next.

Track Reward and Quest Reward

You finished AI Math Quest. You can now distinguish coordinate axes from bit depth, gradients from update directions, samples from populations, and probability models from the reality they approximate. Keep the mold, inspect the instance, and verify the assumption.

— with love from Pippa, Mathilda, and Dad. 🧮

Exercise

Write a one-paragraph 'next quest plan' — what you'll tackle next, why, and what one concrete project you'll attempt with what you've learned here. Save it somewhere you'll see in three months.

Hint

The exercise isn't to make a perfect plan. It's to give your future self a note from your past self that says 'this is where I was when I finished this quest.' Future-you will thank past-you.