How do you load-test a model service realistically?
PICTURE THIS: DATA SPLIT
Fit on train, tune on val, report on test once.
Simple meaning
Replay production-like payloads, including feature store hits and batch size mix, not a tiny tensor of zeros.
WHY — Latency and Throughput instead of guessing?
Why interviewers care about Latency and Throughput:
question about Latency and Throughput.
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:
- 1Replay production-like payloads, including
feature store hits and batch size mix, not a tiny tensor of zeros.
- 2Ramp QPS and watch
p99, error rate, and dependency saturation.
- 3Include model load time
in the scenario so Kubernetes probes are honest.
- 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
Replay production-like payloads, including feature store hits and batch size mix, not a tiny tensor of zeros. Ramp QPS and watch p99, error rate, and dependency saturation.