Finale · One prompt

Replay the same request at system and GPU zoom levels.

Final replay · 12 stages

One prompt, end to end

Follow one request through model readiness, inference, GPU execution, and streamed output without changing timelines.

System spine

Switch zoom levels while preserving the exact replay stage.

Stage 1 of 12

System view follows routing, readiness, scheduling, inference phases, and response streaming.

Core replay

One prompt: system view

one scenario · two synchronized views
PromptHow do GPUs work?
Illustrative responseGPUs run many calculations in parallel.
ClientHow do GPUs work?
Gateway + schedulernot submitted
Model workercoldartifact available
GPUready
Streamed response

No output token yet

waiting
Stage 1 of 12Model artifact exists

Training has already produced a versioned model artifact.

Step 1 of 12
Release outcome

Models, serving systems, and GPU internals are three zoom levels of one token-producing process.