AWS doubles down on custom LLM software with features aimed at simplifying the model creation process

On the heels of the announcement of Nova Forge, a service for training custom Nova AI models, Amazon Web Services (AWS) has announced more tools for enterprise customers to create their own frontier models.

AWS announced new capabilities in Amazon Bedrock and Amazon SageMaker AI at the AWS re:Invent conference on Wednesday. These new capabilities are designed to make it easier for developers to create and fine-tune custom large language models (LLMs).

The cloud provider offers serverless model customization in SageMaker, which allows developers to start building a model without having to think about compute resources or infrastructure, Ankur Mehrotra, general manager of AI platforms at AWS, said in an interview with TechCrunch.

To access serverless form building capabilities, developers can either follow a point-and-click self-guided path or an agent-led experience where they can prompt SageMaker using natural language. The agent-led feature is turned on in the preview.

“If you’re a healthcare customer and you want the model to be able to better understand some medical terms, you can simply point to SageMaker AI, and if you have labeled data, select Technology and then turn off SageMaker, and [it] “It improves the model,” Mehrotra said.

This capability is available to customize Amazon’s Nova models and some open source models (those with publicly available model weights), including DeepSeek and Meta’s Llama.

AWS is also launching Reinforcement Fine-Tuning in Bedrock that allows developers to choose either a reward function or a pre-defined workflow, and Bedrock will automate the model customization process from start to finish.

Frontier LLM software—that is, more advanced AI models—and model customization appear to be an area of ​​focus for AWS at this year’s conference.

AWS announced Nova Forge, a service where AWS will build custom Nova models for its enterprise customers for $100,000 per year, during AWS CEO Matt Jarman’s keynote on Tuesday.

“A lot of our clients ask, ‘If my competitors have access to the same model, how can I differentiate myself?’” Mehrotra said. “‘How can I create unique solutions that optimize, optimize my brand, my data, my use case, and how I differentiate myself?’ And what we found is that the key to solving this problem is the ability to create custom models.

AWS has not yet gained a large user base for its AI models. A July survey by Menlo Ventures found that companies overwhelmingly preferred Anthropic, OpenAI, and Gemini over other models. However, the ability to customize and fine-tune these LLM programs can begin to give AWS a competitive advantage.

Follow all of TechCrunch’s coverage of the annual enterprise technology event here, and see all the announcements you may have missed so far here.