RESOURCES
Get to know
the experiment.
Explore the workflow, understand the demo’s scope, and find answers before entering the lab.
Inside the agent team.
Explore the three specialists, the arbiter, and the evidence behind each decision.
Read the product guide ↗WORKFLOW GUIDEFrom tuning to sharing.
See how individual optimization connects to a joint trial on shared hardware.
Explore the workflows ↗RUNS AND RECORDSFollow the decisions.
Open demo for the live loop, or inspect saved trial records in the dashboard. The dashboard is not a live feed from the notebook.
Enter the lab ↗A FEW USEFUL DETAILS
Before you
enter the lab.
Do I need a GPU to try the demo?
The offline replay and local simulator need no GPU or API keys. Live demo needs a compatible GPU, an OpenAI key for investigators, and a W&B key and Weave project. Clone the notebook before entering keys. GPU time and hosted model calls can cost money.
Are the results hardware benchmarks?
The recorded Qwen2.5-72B run measured 19.77% lower p95 latency on eight warmed, repeated tasks, with 8/8 answer checks passing. Simulator numbers are modeled, not measured GPU results.
Does Sera always recommend sharing?
No. A memory fit is followed by a joint trial. Each model must pass its own performance and quality gates; a failed gate can lead to a revert.
What makes this different from a parameter sweep?
Investigators read Weave evidence, review each other’s proposals, and adapt after each trial. The full run demonstrates this loop. A general search-efficiency advantage over grid or random search is not established.
SEE THE LOOP IN ACTION