Design an LLM Agent System That Automatically Resolves Jira Tickets and Opens Pull Requests
Company: Datadog
Role: Software Engineer
Category: ML System Design
Difficulty: medium
Interview Round: Technical Screen
Overview: This ML system design question evaluates the ability to architect an autonomous LLM-agent pipeline spanning retrieval-augmented generation, tool integration, and sandboxed code execution. It tests conceptual understanding of how to ground agent reasoning in real codebases, structure permissioned tool access, and safely isolate execution environments — core competencies for senior roles building AI-powered developer tooling.