Back to problems

ML Coding from Scratch (Regression / Markov / Facility)

Algorithm · Uber · Hard

Requirements The interviewer may select one of these problem families: Linear / logistic regression — write closed-form linear regression with the normal equations, followed by logistic regression trained with mini-batch gradient descent. You should derive the gradient on the whiteboard before implementing it. The detailed specification appears in ### Linear and logistic regression from scratch. Markov-chain text generator — implement two functions:…

Checking your access…