Skip to main content
T1 Large companies and unicorns Xiaomi New grad Round 1

Xiaomi AI Infra campus hire round 1

Xiaomi AI Infra campus-hire round-1 interview questions covering inference optimization

Inference optimization

Round 1

Inference optimization

  1. Introduce the core technical details of Fastllm.cpp.
  2. Role and design of PagedAttention.
  3. Algorithm flow of FlashAttention.
  4. Basic principles of model quantization and common schemes.

Round 2

Project experience

  1. Deep dive into projects and technical details.