ML
ML Interview Roadmap
← Study plan·Generative AI

Day 76 of 133

Reasoning models (o1/o3, Claude thinking) + DSA Intervals

Test-time compute; CoT vs ToT vs self-consistency.

DSA · NeetCode Intervals

  • Insert IntervalDSA · Intervals

    Interview questions to prep

    1. Do you sort by start or by end? Defend the choice based on the invariant you need.
    2. Walk through merge / overlap detection: what's your condition for 'overlapping'?
    3. How does complexity break down: O(n log n) sort + O(n) sweep — can you do better in any case?
  • Merge IntervalsDSA · Intervals

    Interview questions to prep

    1. Why sort by start (not end) for THIS problem — and contrast with non-overlapping-intervals.
    2. How does this scale if intervals stream in unsorted?

GenAI · Reasoning models (o1, o3, thinking)

  • o1 / o3: long reasoning, RL on reasoning tracesGenerative AIOpenAI

    Interview questions to prep

    1. How do reasoning models like o1 differ from standard chat LLMs?
    2. When are reasoning models worth their cost, and when are they overkill?
  • Claude extended thinking & test-time computeGenerative AIAnthropic

    Interview questions to prep

    1. What does test-time compute scaling buy you over training-time compute?
    2. When does extended thinking actually hurt — what tasks does it slow down without quality gains?
    3. Why is longer reasoning output not the same thing as better reasoning?
    4. How would you compare self-consistency against selecting generations by a reasoning-quality signal such as deep-thinking ratio?
    5. If you could reject weak reasoning samples after the first 50 tokens, how would that change inference cost and quality?
  • Tree of Thoughts, self-consistencyGenerative AIYao et al.

    Interview questions to prep

    1. Compare CoT vs ToT vs self-consistency — when does each help?
    2. Why does self-consistency improve reasoning, and when does it fail?

References & further reading