Skip to content
Skills
Problems
Questions
Drills
Roles
Skills
Interview
Contest
Leaderboard
Problems
Questions
Drills
Roles
Skills
Interview
Contest
Leaderboard
Home
/
MLOps Engineer
/
Questions
/
LLM Serving
MLOps Engineer
LLM Serving questions
7 questions on LLM Serving for mlops engineers, easiest first.
Easy
1
Memory for 7B parameters in fp16
easy
Medium
2
KV cache size for one long request
medium
Serving many fine-tuned variants
medium
Hard
4
Why LLM servers use continuous batching
hard
Where to use tensor parallelism
hard
KV cache with grouped-query attention
hard
Semantic cache returns wrong answers
hard
