Skip to main content
T0 Big tech Baidu Round 1

Baidu AI Infra first round (2)

Real Baidu AI Infra first-round interview questions covering inference optimization and operator optimization

Inference optimization Operator optimization

Round 1

Project deep dive

  1. Introduce the project’s core content and technical design

Theory

  1. How GEMM is implemented and optimized in CUDA

Live coding

  1. Implement a CUDA vector-add kernel

Round 2

Theory

  1. What common optimizations are used when deploying LLM inference?
  2. How do you convert between Float32 and INT8 (model quantization)?

Live coding

  1. Longest common subsequence (LeetCode 1143)