Skip to main content
T5 Others Lenovo Internship Round 1

Lenovo AI Infra internship first round

Real Lenovo AI Infra internship first-round interview questions covering AI Infra and related topics

Fundamentals

  1. What attention optimization does DeepSeek R1 use?
  2. Introduce the overall structure of Transformer.
  3. How have mainstream LLM architectures changed compared with earlier ones?
  4. What improvements exist for GRPO? Do you know GSPO?
  5. What common KV Cache optimization methods exist?

Project experience

  1. Introduce the project (background, highlights, ownership, problems and solutions, why those tech choices, training data volume, GPU count and training time, team makeup).
  2. For some technical point in the internship project, do you think there is a better approach? What are the pros and cons of that approach?

Coding

  1. Hand-implement Multi-Head Attention, and say at which step the Masked matrix should be introduced.