chini-009-video-upload

Video Upload Pipeline

Accept large uploads. Transcode in the background. Survive a worker meltdown.

Source: Classic system-design interview corpus (YouTube / TikTok upload + transcode)

Prompt

Design a video upload and transcode pipeline.

Functional:
- POST /upload accepts a video file and returns immediately with an upload id.
- A background pipeline transcodes the source into multiple resolutions (480p, 720p, 1080p).
- When transcoding finishes, the asset is published and watchable.

Non-functional:
- A 4x burst of uploads must not break the accept path; backlog is acceptable, dropped uploads are not.
- If the transcode worker pool partially fails, in-flight jobs should retry and complete.
- Upload acceptance must not block on transcode completion (decoupled paths).

Return a Chinilla CanvasState. Expect a queue between accept and workers, retry behavior, and storage for the source + variants.

Constraints

Max components: 13
Required behaviors: queue, retry, storage
Monthly budget: $1500

Stress scenarios

Baseline uploads

baseline

Normal upload + transcode volume.

4x upload burst

spike

Creator event causes 4x uploads. Accept path must hold.

Worker failure

outage

A transcode worker dies. Jobs must retry and complete elsewhere.

Slow transcode

latency

Transcode latency triples. Pipeline must absorb without dropping uploads.

Pass criteria (overall)

Min stability score: 65
Max drop rate: 6.0%
Min delivery rate: 88.0%
Max errors: 5

Submit your run

Submissions go through the chini-bench CLI. It calls your model with your key, scores the result locally, and posts to the leaderboard. Nothing leaves your machine except the canvas it produces.

End-to-end:

pip install git+https://github.com/collapseindex/chini-bench-cli.git
export OPENROUTER_API_KEY=...

chini-bench run chini-009-video-upload \
  --provider openrouter --model google/gemini-2.0-flash-001 \
  --as alice --x alice --linkedin alice-builds

Or inspect the prompt first:

chini-bench prompt chini-009-video-upload

Providers: openai · anthropic · google · openrouter · ollama

Leaderboard

Rank	Submitter	Model	Score	Stability	Delivery	Design	Pass	Links
#1	alex default	A anthropic/claude-sonnet-4.6	83	72.0	88.0	100.0	✗	X
#2	alex default	G google/gemini-3.1-pro-preview	83	72.0	88.0	100.0	✗	X
#3	alex default	X x-ai/grok-4.20	82	80.0	75.0	100.0	✗	X
#4	alex default	O openai/gpt-5.4	81	63.0	100.0	100.0	✗	X

Per-scenario breakdown of the top run

Scenario	Health	Drop rate	Delivered	Pass
baseline	70.0	0.0%	1400	✓
upload-burst	70.0	0.0%	6400	✓
worker-failure	79.0	0.0%	12	✗
slow-transcode	70.0	0.0%	350	✓

How is this scored? →