What should an ML CI pipeline test?
PICTURE THIS: DATA SPLIT
Fit on train, tune on val, report on test once.
Simple meaning
Training code unit tests, data schema checks, and a smoke train on a tiny sample.
WHY — CI instead of guessing?
Why interviewers care about CI:
who only read docs from people who shipped.
and tied to MLOps work.
Name the idea, why it exists, then one short example.
End with when you use it and one common pitfall.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1Training code unit tests,
data schema checks, and a smoke train on a tiny sample.
- 2Full GPU train may
be nightly.
- 3How it works
Broken preprocessing should fail fast.
- 4Give an example
One tiny concrete case you can say aloud.
- 5Common mistake
What juniors usually get wrong.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
Training code unit tests, data schema checks, and a smoke train on a tiny sample. Full GPU train may be nightly.