Workspace preview

Discuss a pilot ↗

INTERFACE ONLY

No code execution or AI assistance. Explore freely.

Role

SeeTheWork OSInvestigate a suspicious model improvementPreview · no time limit

Code Studio

Open fileExample file

1# Keep each customer in one split.
2splitter = GroupShuffleSplit(
3    n_splits=1, test_size=0.2, random_state=42
4)
5train, holdout = next(splitter.split(
6    features, target, groups=customer_id
7))
8assert not set(customer_id[train]) & set(
9    customer_id[holdout]
10)
11pipeline.fit(features[train], target[train])
12prediction = pipeline.predict(features[holdout])

OUTPUT

Code execution is disabled in this preview.

You can browse and edit the example files. Changes stay in this tab until you reset or leave.

Team Chat

Scenario context

PREWRITTEN SCENARIO MESSAGE

ML lead

The new score looks unusually strong. Can we trust the comparison with the baseline?

A place to clarify the work.

In a live simulation, candidates can ask role-specific coworkers questions here.

Messaging and AI replies are disabled in this preview.

Browse files. Arrange the workspace. Get a feel for the candidate experience.