Skip to main content
T1 Large companies and unicorns JD.com New grad Round 1

JD.com AI Infra campus recruiting first round

Real JD.com AI Infra campus recruiting first-round interview questions covering inference optimization

Inference optimization

Inference optimization

  1. What common methods exist for large-model inference optimization? Cover several dimensions.
  2. What is the core principle and design idea of PagedAttention?
  3. Why can FlashAttention accelerate? What does its compute process look like?
  4. In PD disaggregation, how is the scheduling queue implemented?
  5. Does Chunked Prefill mean Prefill and Decode of the same request interleaved, or insertion across different requests?

Fundamentals

  1. How is C++ polymorphism implemented?
  2. What is the concrete mechanism of virtual-function polymorphism? How is the order of functions in the vtable determined?

Project experience

  1. Deep dive on project experience.

Coding

  1. Implement quicksort.