[shipmill-romamo]

The Nebius x NVIDIA hackathon asks the README to say where Token Factory accelerated the workflow and which other Nebius tools or services were used. The Nemotron section covered how the model is used but not those two points.

Two bullets under "How Nemotron and Token Factory are used":

• Token Factory sped up the work: OpenAI-compatible tool calling worked on the first call for all four Nemotron models, so the 40-run evaluation needed no GPUs or deployment of our own
• Other Nebius services: Token Factory only at runtime; the evaluation is packaged for a Serverless Job (its Dockerfile), not run there because AI Cloud was not activated for our tenant

README only, so no changelog fragment, like earlier README-only changes.

🤖 Generated with Claude Code