Improve
SUPERVISED FINE-TUNING
Serverless SFT
Fine-tune LLMs on examples of the behavior you want while CoreWeave manages elastic GPU capacity and distributed training. Move between supervised fine-tuning (SFT) and reinforcement learning (RL) in one workflow without shuttling artifacts across systems.
SERVERLESS SFT IS PART OF COREWEAVE FORGE
Turn curated examples into better model behavior
CoreWeave Forge connects the data, training, evaluation, and inference behind your AI application. Serverless SFT turns examples of the behavior you want into a task-specific model that you can test, evaluate, and continue improving.
FAQS
Frequently asked questions
What is Serverless SFT?
When should I use Serverless SFT?
What kind of training data does Serverless SFT use?
Do I need to provision a training cluster?
What does Serverless SFT train?
Can I move from SFT to reinforcement learning?
Where are trained models stored?
How do I test a trained model?
Builder Resource Center: Learn from every run
Explore demos, code, and technical resources for every stage of the AI loop. Learn how researchers, developers, and CoreWeave engineers build, observe, evaluate, and improve AI systems—and put those insights to work.
GET STARTED
Teach your model the behavior you want
Bring your curated examples. CoreWeave runs the GPUs and distributed training while you keep control of the training loop. Test the checkpoint with Serverless Inference, then continue with Serverless RL when you're ready to optimize for outcomes.
