Amazon’s Bedrock and SageMaker AI get more customizable.
Amazon’s continued focus on offering AI agents through AWS now includes Reinforcement Fine Tuning of AI models built on Amazon’s Nova 2 Lite in Bedrock. The company also announced new “serverless” model customization with SageMaker AI that can provide compute based on the model size with Amazon’s own Nova, OpenAI’s GPT-OSS, Meta’s Llama, and others.
TL;DR
- Amazon Bedrock now supports Reinforcement Fine Tuning of AI models.
- This fine-tuning can be applied to AI models built on Amazon's Nova 2 Lite.
- Amazon SageMaker AI introduces serverless model customization.
- Serverless customization provides compute based on model size.
- Supported models include Amazon's Nova, OpenAI's GPT-OSS, and Meta's Llama.