LLM Application Engineering: RAG vs Fine-Tuning, Context Limits, State and Prompt Injection
Company: Mercor
Role: Software Engineer
Category: Machine Learning
Difficulty: medium
Interview Round: Technical Screen
Overview: Four rapid-fire questions on building applications with large language models: how retrieval-augmented generation relates to fine-tuning, what to do when requests exceed the context window, how to make stateless model calls stateful, and how to defend a tool-using agent against prompt injection. It tests practical, production-level judgment.
Read the full Mercor Software Engineer interview experience this question came from