![](https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1/files/2023/11/Bratin-clean.jpg)
![](https://d15shllkswkct0.cloudfront.net/wp-content/blogs.dir/1/files/2023/11/Bratin-clean.jpg)
Artificial intelligence is running at such a speed that companies not keeping pace are immediately left behind. Compared to other players, such as OpenAI Inc. and Microsoft Corp., there’s been a long-held notion that Amazon Web Services Inc. is behind the AI innovation curve.
That couldn’t be further from the truth, however, as AI underpins many of its current and future offerings, according to Bratin Saha (pictured), vice president and general manager of AI and machine learning at Amazon Web Services Inc. The company’s heritage is encapsulated in its proprietary three-layer stack, which starts at the infrastructure level.
“It starts with the infrastructure where we have our GPUs, but we also have our custom processors, Trainium and Inferentia,” Saha said. “You get significant performance benefits and cost performance benefits from those. And then you have SageMaker, which provides you the end-to-end software infrastructure for building and training.”
Saha spoke with theCUBE industry analyst John Furrier at the “Supercloud 5: The Battle for AI Supremacy” event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed Amazon’s deep-rooted history in AI, the three-layer stack and the unveiling of Amazon Q.
SageMaker takes the spotlight as the end-to-end software infrastructure for building and training machine learning models, catering to those who wish to deploy their own models. On the other hand, Bedrock, offering a myriad of state-of-the-art models, stands as a testament to Amazon’s commitment to providing choice and innovation as a key differentiator, according to Saha.
“Then you have customers who say, ‘No, I just want to use models that are provided by Amazon or others’ — then we have Bedrock,” he said. “That provides you the most choice that there is out there of state-of-the-art models along with other capabilities we talked about today, like [retrieval augmented generation] knowledge bases and so on.”
The three-layer stack culminates in a suite of applications, including the just-unveiled Amazon Q, poised to revolutionize data interaction for business users, contact centers and AWS enthusiasts. In a nutshell, Q is an AI-powered assistant that’s been trained on 17 years’ worth of AWS knowledge and helps users with unique responses across a myriad of topics and use cases. Uniquely, Q has been designed to assist developers as well as non-tech users, according to Saha.
“We think it’s going to transform the way employees interact with the data,” he explained. “You have Q for business users, you have Q for the contact center, you have Q for AWS. I think it’s a really good set, and it just shows that we’ll keep on innovating. We had HealthScribe, we had CodeWhisperer, and now we have Q and we just keep innovating.”
In the middle of that three-layer stack is the LLM foundation model layer, according to Saha. Synthetic data has emerged as a key player, addressing scenarios where obtaining real data proves challenging. There’s a critical role of synthetic data in scenarios like defect detection, shedding light on its potential to create a new era of digital twins and simulations.
Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the “Supercloud 5: The Battle for AI Supremacy” event:
THANK YOU