Loading…
The maths behind AI, without the fear: vectors and embeddings, matrices, derivatives and gradient descent, statistics, probability and Bayes, LLM sampling and evaluation metrics — all in NumPy.
Seventeen pages of the maths every AI engineer uses, taught with code instead of proofs. Vectors, dot products and cosine similarity (how semantic search works); matrices and neural-network layers; derivatives, gradient descent and loss functions (how models learn); averages, spread, distributions, correlation and confidence intervals (how to read data and results honestly); probability, Bayes' theorem and how a language model picks its next word; and the precision, recall and F1 you need to judge any model.
Every formula is real NumPy code with its real output, explained in English and Bangla, with exercises on every page and three quizzes from easy to hard.
6 chapters · 56 min total
Watch it taught, then check what stuck.