Design a hybrid system that precomputes 95 percent of scores but handles 5 percent of cold-start users online.
PICTURE THIS: HOW TO EXPLAIN IT
Simple meaning
A batch job would score known entities into a KV store
WHY — Batch vs Realtime instead of guessing?
Why interviewers care about Batch vs Realtime:
separate people 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:
- 1A batch job would
score known entities into a KV store
- 2the API would lookup
first and only call the model for cache misses or stale TTLs.
- 3Cold-start users would use
a lighter model with fewer features to protect p99.
- 4Metrics must split cache-hit
versus miss latency so you do not hide a broken online path.
- 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
A batch job would score known entities into a KV store the API would lookup first and only call the model for cache misses or stale TTLs.