Design Product Discovery Inside Video

1. Problem Statement

2. Architecture Discussion Map

Use this as one discussion aid, not a single correct answer. Your design should follow from the requirements, scale, and trade-offs you establish.

graph LR
    A["Design Product Discovery Inside Video"]
    A --> F1["Frame sampling and scene detection"]
    A --> F2["Offline versus online inference"]
    A --> F3["Product retrieval and confidence thresholds"]
    A --> F4["Human review and catalog freshness"]
    A --> F5["Time-aligned low-latency serving"]

3. Key Focus Areas

• 1
  Frame sampling and scene detection
• 2
  Offline versus online inference
• 3
  Product retrieval and confidence thresholds
• 4
  Human review and catalog freshness
• 5
  Time-aligned low-latency serving

4. What the Interviewer Looks For

The interviewer probes these areas, and the report grades the interview against them.

• Processing mode and compute budget
• Detection and catalog matching
• Quality control and human review
• Time-aligned serving and playback isolation
• Catalog freshness, policy, and operations

5. What Strong Candidates Should Demonstrate

• Build an asynchronous video-to-frame-to-product pipeline.
• Join uncertain visual matches with a searchable product catalog.
• Serve time-aligned purchase links without blocking video playback.