Skip to main content
T5 Others Shopee Internship

Shopee AI Infra internship (2)

Real Shopee AI Infra internship interview questions covering inference optimization and related topics

Inference optimization

Inference optimization

  1. Core principles and design ideas of vLLM
  2. Introduce vLLM, model quantization, and KV Cache optimization tricks
  3. How to improve multi-turn dialogue ability of a model
  4. What is the difference between “packing” form and “multi-turn dialogue” form

RAG

  1. How do you evaluate RAG app quality, and which metrics did you use
  2. How do you train a generation model for RAG scenarios
  3. How to construct CoT (Chain-of-Thought) training data

Coding

  1. Edit distance (LeetCode 72)