ML
ML Interview Roadmap
← Study plan·MLOps

Day 62 of 133

Deployment patterns: batch / real-time / streaming / edge + DSA 1-D DP

Latency budgets, async inference, when to push to device.

DSA · NeetCode 1-D DP

  • Longest Palindromic SubstringDSA · 1-D DP

    Interview questions to prep

    1. Compare expand-around-center (O(n²) time, O(1) space) vs Manacher's (O(n)).
    2. Why is DP O(n²) time AND O(n²) space — and does it actually beat expand-around-center in practice?
  • Palindromic SubstringsDSA · 1-D DP

    Interview questions to prep

    1. State the DP: define the state, the transition, and the base case explicitly.
    2. Top-down (memoized recursion) vs bottom-up (tabulation) — which is more natural here, and why?
    3. Can you space-optimize from O(n) to O(1)? Show the rolling-window trick.

MLOps · Deployment patterns

  • Batch vs real-time vs streaming inferenceMLOpsChip Huyen

    Interview questions to prep

    1. Compare batch, real-time, and streaming inference — when do you reach for each?
    2. What's the latency budget for an ad CTR model vs a churn model?
  • Edge deployment & on-device MLMLOpsPyTorch Mobile

    Interview questions to prep

    1. When does on-device ML beat cloud inference, and what are the constraints?
    2. What model-side techniques (quantization, pruning, distillation) actually move the needle for mobile?
  • Async inference & queuesMLOpsAWS

    Interview questions to prep

    1. When would you use async inference behind a queue?
    2. How do you size a queue worker pool for spiky inference traffic without over-provisioning?

References & further reading