Nvidia and Dell recently teamed up to create Project Helix, a project that promises to bring generative artificial intelligence capabilities to organizations. By allowing organizations to use their own data to build new generative AI-based models or modify existing ones, it will make it easier to get started with generative AI. Because this work is done within the corporate structure, Project Helix can also help mitigate the intellectual property leakage that many companies fear.
Also read: Samsung and Naver to Develop AI Replica to Compete with ChatGPT
The product focuses on the development of full-stack, in-house generative artificial intelligence solutions that enable organizations to build new or modify existing generative artificial intelligence base models using their own data. As the solution is explicitly designed to make it easier to get started with generative AI, it can help businesses avoid common pitfalls and cost overruns when implementing generative AI projects.
“Project Helix gives enterprises purpose-built AI models to more quickly and securely gain value from the immense amounts of data underused today,” said Jeff Clarke, vice chairman and co-chief operating officer, Dell Technologies. “With highly scalable and efficient infrastructure, enterprises can create a new wave of generative AI solutions that can reinvent their industries.”
As an artificial intelligence company, Nvidia unveiled a complete set of software related to generative artificial intelligence, including industry-specific software foundational models and enterprise-focused development tools, during its recent GTC presentation in March. Many people were startled by the NeMo Large Language Model (LLM) structure and the NeMo Guardrails for screening out unwanted topics. These models have been tailored for use with Nvidia hardware.
Project Helix is also a joint venture between Nvidia and Dell, which has packaged Nvidia’s enterprise AI software into a range of PowerEdge server systems. In addition, Dell offers many storage options tailored for AI workloads through its PowerScale and ECS enterprise object storage lines. As a result, organizations can begin building or adapting generative AI models with a complete solution. Prospective customers can begin the process using Nvidia’s base model or by selecting an open-source model from Hugging Face (or solutions from other technology providers).
Project Helix elements will be available through traditional channels as well as APEX flexible consumption options, with services beginning in July 2023.