Skip to main content
T1 Large companies and unicorns vivo New grad

vivo AI Infra campus recruiting

Real vivo AI Infra campus recruiting interview questions covering operator optimization

Operator optimization

Project deep dive

  1. Introduce related project experience
  2. Overall backend architecture and key implementation details

Theory

  1. Core ideas of HalideIR and how Schedule primitives work
  2. Besides TVM, which mainstream DL compilers exist (XLA, TensorRT, and so on)? How do they compare?
  3. How does TVM do graph-level versus operator-level optimization? What support does the target hardware give?
  4. Design ideas for auto-tuning? Pros and cons versus pure manual optimization?