Skip to main content
T2 AI unicorns MiniMax Internship Round 1

MiniMax AI Infra internship first round (2)

Real MiniMax AI Infra internship first-round interview questions covering training optimization

Training optimization

Project questions

  1. Deep discussion of your project

Theory

  1. How does MoE keep training efficient as parameter scale grows?
  2. How do SFT and RLHF differ in optimization target? Why is extra RLHF usually needed after SFT?
  3. What objective does PPO optimize in RLHF? Write the objective and explain each part
  4. Briefly introduce the LLM architecture you know best and its traits
  5. How does MoE routing work? What causes uneven expert load?
  6. When expert utilization is poor, what routing fixes are practical?

Coding

  1. Write a program for the sliding-window maximum