The machine learning and deep learning basics that still get asked
AI engineer loops open with fundamentals and come back to them whenever you claim LLM experience. This lesson is a study guide to the ones that recur: diagnosing bias and variance, regularisation, splits and leakage, metrics under imbalance, losses as likelihoods, optimisers and schedules, backprop, normalisation and embeddings, with the numbers and the wrong answers for each.
What you'll be able to do
- Diagnose underfitting or overfitting from a train and validation curve and name the next change to make
- Explain cross-entropy, weight decay and AdamW as consequences of likelihood and of how Adam scales its steps
- Choose a metric and a validation split for a rare-positive or time-ordered problem, and name the leaks to audit for
- Describe how backprop, residual connections, normalisation and initialisation keep a deep network trainable
- Pick the right similarity measure for an embedding model and say what breaks when the index metric does not match
Before this: the-ai-engineer-loop
Log in to read the rest of this lesson
AI engineer loops open with fundamentals and come back to them whenever you claim LLM experience. This lesson is a study guide to the ones that recur: diagnosing bias and variance, regularisation, splits and leakage, metrics under imbalance, losses as likelihoods, optimisers and schedules, backprop, normalisation and embeddings, with the numbers and the wrong answers for each. This walkthrough runs about 18 minutes, with runnable code you can edit and re-run as you read. A new account gets 7 days of Pro free, no card required.
Log in to continueThe first module of every track is free to read on the web: see what's open in AI interviews.