Design an Agentic Ticket Support System
Company: Giga
Role: Software Engineer
Category: ML System Design
Difficulty: medium
Interview Round: Onsite
Design an AI-powered ticket support system.
The system must integrate with two existing external systems:
1. **Ticket system**: Displays all support tickets in a web UI, but it does not provide an API. The AI agent must use a browser-based automation approach to read ticket data.
2. **User system**: Provides an API that can be used to retrieve user information. This user data should be used to generate personalized chat responses.
Design the end-to-end architecture. Explain how tickets are discovered, how the agent reads ticket content from the browser, how it retrieves user context, how it generates responses, and how the system ensures reliability, security, observability, and human review when needed.
Quick Answer: This question evaluates a candidate's ability to design ML-driven production systems that combine agentic AI, browser-based automation for services without an API, and external user-data APIs to produce personalized support responses.