Moderate Train vs Serve Question 73 of 221

How do you decide CPU versus GPU for inference?

MLOps track · Speak this in 60–90 seconds · Faridabad & Delhi NCR

PICTURE THIS: DATA SPLIT

Train 70%Val 15%Test 15%

Fit on train, tune on val, report on test once.

Simple meaning

I look at model size, p99 budget, QPS, and cost.

1

WHY — Train vs Serve instead of guessing?

Why interviewers care about Train vs Serve:

They want a clean

contrast on Train vs Serve, not two memorised paragraphs.

Say what changes for

the developer, then one case where picking wrong hurts.

Stay structured

Name the idea, why it exists, then one short example.

Close cleanly

End with when you use it and one common pitfall.

2

STEPS — What happens step by step?

Before you speak the answer, walk the interviewer through these steps:

  1. 1
    I look at model

    size, p99 budget, QPS, and cost.

  2. 2
    Small tabular models almost

    always stay on CPU

  3. 3
    large transformers may need

    GPU or a compiled runtime.

  4. 4
    I would benchmark both

    with production-like batch sizes because GPU helps only if the hardware stays busy.

  5. 5
    Common mistake

    What juniors usually get wrong.

  6. 6
    Close

    When you pick this over the alternative.

3

EXAMPLE — See it in action

Here's a short line you can speak, broken into clear beats:

Say this line
“Small tabular models almost always stay on CPU”
Break into beats
Smalltabularmodelsalmostalwaysstay
Speaking order
2987408337471632900

Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.

Key takeaway

I look at model size, p99 budget, QPS, and cost. Small tabular models almost always stay on CPU

Chat with us