The ML Coding Round

This round is for machine learning and research engineering candidates. If you're on a backend, full-stack, platform or forward deployed track, you won't see it, and your hours are better spent on the practical coding round. On the ML track it's the technical phone screen, and candidates who've taken it describe the same flavor going back years. You do hands-on NumPy work over distributions, sampling and simple statistics, then answer questions about your past ML work.

It runs in a notebook or a shared editor and feels like a normal day on an ML team, with sampling, scoring and data cleaning. The math stays light. Most of your time goes to array shapes, messy records and explaining each step.

In the real round you'd draw random samples or load files from disk. Here the inputs come in as plain lists, so each answer is fixed and can be checked. Write each one as a small function first. Then talk through the numpy or pandas version you'd use at full size, since interviewers on this track want to hear that you know how the code would change with real data.

Scale AIPremium lesson · Coding round

The 3 Scale AI problems are below.

1. 1A Premium questionCoding · Medium · Frequently asked
2. 2A Premium questionCoding · Medium · Frequently asked
3. 3A Premium questionCoding · Medium

Each with hints, the follow-up and a tested solution.

Get the Scale AI questions

$49 a year · all 100 companies · cancel anytime

From 20,000+ candidate reports, forum threads behind logins and write-ups in Chinese.

Read a free Scale AI lesson firstAlready a member? Sign in