How do you decide when to scale model serving pods?
PICTURE THIS: DATA SPLIT
Fit on train, tune on val, report on test once.
Simple meaning
Latency SLOs, queue depth, and CPU/GPU utilization.
WHY — Cost instead of guessing?
Why interviewers care about Cost:
question about Cost.
trade-offs, and what you would actually do on a MLOps project - not buzzwords.
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:
- 1Latency SLOs, queue depth,
and CPU/GPU utilization.
- 2Scaling on CPU alone
misses GPU saturation.
- 3How it works
Load tests set the thresholds.
- 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
Latency SLOs, queue depth, and CPU/GPU utilization. Scaling on CPU alone misses GPU saturation.