Skip to main content

Assessment Studio

Evaluate a model before launch

Decide whether a churn model is ready to launch, and say what you'd check first.

Discipline
Data, AI & Intelligence
Time
75 minutes, from when you start
AI tools
AI tools allowed (declare how you used them)
A pass proves
Machine Learning, Model Evaluation, Python

What you hand in

A written review, as text or a link to a notebook or document.

How it's marked

A person scores each line from 0 to 4. Lines carry different weight. You pass with 70% or more of the weighted total and no line at 0.

What's markedWeight
Metric judgementSees that 94% accuracy is close to always predicting no churn.×3
LeakageSpots the time leakage in the tickets column and explains the fix.×3
Evaluation planSensible metrics, baselines and a safe rollout test.×2
CommunicationClear enough for a teammate to act on without a meeting.×1

Start

Assessments are for verified members. Sign in or join NXG.