Dell and NVIDIA joined forces to put generative AI into the hands of Dell’s software-as-a-service customers.
Project Helix will be Dell’s first foray into artificial intelligence for its edge software service, Dell Technologies Senior Vice President of Product Marketing Varun Chhabra announced as part of a preview briefing for the Dell Technologies World conference. The service lets organizations build and deploy generative AI and includes security and trustworthiness considerations to reduce inaccuracies or vulnerabilities.
What is Dell’s Project Helix?
Project Helix is an early look at an upcoming Dell product that will assist organizations in running generative AI. It spans the lifecycle of the AI creation process, including infrastructure provisioning, modeling, training, fine-tuning, application development and deployment.
“I don’t think any of us have seen in the last 20 years the kind of game changer generative AI is,” said Sam Grocott, Dell Technologies’ senior vice president of product marketing.
In Project Helix, Dell and NVIDIA created a partnership to deliver full-stack, scalable solutions using Dell’s compute, storage and software infrastructure and NVIDIA’s accelerators, AI software and foundational models and experience delivering generative AI.
Dell will provide servers, including the AI-optimized Dell PowerEdge servers; NVIDIA’s H100 Tensor Core GPUs and NVIDIA Networking will support the workloads. From there, customers can use Dell CloudIQ software for observability and NVIDIA AI Enterprise tools for management throughout the AI’s lifecycle.
SEE: Dig deep into observability needs for AI-as-a-service
“Enterprises are now actively looking at how they can put generative AI to use for their enterprises, whether it’s to improve back-end performance or improve customer experience,” Chhabra said.
However, customers looking for generative AI want to be certain the output will be accurate and appropriate and will not expose their proprietary information to the outside world or put them in breach of privacy laws and regulations, he added. Dell aims to solve this problem in part by focusing on purpose-built generative AI models rather than more general ones like ChatGPT. Customers will be able to create or tune their own AI for their own domains.
Another challenge customers face is training people to use something as novel as generative AI, Chhabra said. Therefore, Project Helix’s services will include expertise and guidance for customers at all points in the AI lifecycle. Dell indicates it wants to balance agility and time to market with privacy and regulatory compliance, aiming to differentiate its AI through trust as much as performance.
Different solutions based on Project Helix will be released gradually, but more information about the first one is expected to be released in June 2023.
Competitors to Project Helix
Amazon has a machine learning equivalent to Project Helix: SageMaker, which includes its own chat AI, Bedrock. Google Cloud Platform’s AutoML and Vertex AI, as well as Microsoft Azure, are also competitors in this area.
Disclaimer: Dell paid for my airfare, accommodations and some meals for the Dell Technologies World event held May 22-25 in Las Vegas.