[Dataiku Japan Co., Ltd.]
Dataiku expands LLM mesh to enable multi-LLM approaches across cloud and AI player ecosystem
*Dataiku Japan Co., Ltd.*
Press release: August 14, 2024
**
Dataiku expands LLM mesh to enable multi-LLM approaches across cloud and AI player ecosystem
*New Dataiku LLM Registry further strengthens the LLM mesh by adding a layer of governance to assess, document, and frame LLM usage* Dataiku provides a universal AI platform today announced expansions to its LLM mesh ecosystem. It provides secure access to thousands of large-scale language model (LLM) gateways, allowing data and analytics teams to adopt a multi-LLM strategy to build and deploy generative AI-driven solutions at scale. It will look like this. With the LLM Registry, Dataiku also fills a critical governance gap to enable regulatory compliance and effective management of LLM technology across organizations. This allows CIOs and their teams to see, document, and streamline which LLMs should and should not be used across use cases.
In a highly competitive and volatile LLM ecosystem, Dataiku’s LLM Mesh enables organizations to take a multi-LLM approach, allowing them to switch between underlying models and power their applications with generative AI. Now you can. With this expansion, LLM Mesh is now available on Amazon.com.
Web Services (AWS), Databricks, Google
Supports numerous LLM players, including 15 major cloud and AI vendors such as Cloud, Snowflake (Arctic), etc.
Florian Douetteau, Founder and CEO of Dataiku, said:
“Our goal is to help our customers stay ahead of the curve by future-proofing their generative AI strategies, while eliminating the risk of relying on a single AI provider for their strategies. , LLM Mesh provides a balanced approach to developing AI applications for literally any generative AI use case that enterprises are looking to implement today for a true multi-LLM strategy. We provide secure access to a thousand diverse models.”
LLM is a subset of generative AI applications, and the reality of LLM use in enterprises is becoming more complex as organizations scale into more advanced applications. A multi-LLM approach is essential to manage cost and performance, consider privacy and security, and meet regulatory requirements. Dataiku’s Universal AI Platform supports this holistic approach, with the help of traditional analytics and machine learning technologies, allowing companies to effectively operate the entire development lifecycle of generative AI applications. .
Nancy Gohring, senior research director for AI at IDC, said: “IDC predicts that there will be many different types of models, each suitable for different tasks and scenarios. We need to have the ability to quickly evaluate and replace models as new models come to market and use cases evolve.”
As the only infrastructure-agnostic vendor in the industry, Dataiku decouples generative AI applications from the service layer and provides guardrails in terms of cost, usage, hallucination, PII, and more. Such choice and flexibility allows Dataiku to scale its unique integrations at the pace the ecosystem evolves, allowing organizations to create generative AI applications that deliver business value rather than ramping up new LLM connections. Be able to build and deploy.
Reference materials
– O’Reilly’s latest book on the challenges of using LLM in enterprises, *The LLM Mesh: A Practical Guide to
*Chapter 1* of Using Generative AI in the Enterprise*
Please see
Please take a look.
About Dataiku
Dataiku is the platform for universal AI, from modern analytics to generative AI, enabling you to integrate technology, teams, and businesses to bring intelligence to your daily work. Together, they enable companies of all sizes and in all industries to design, develop, and deploy new AI-powered capabilities. Dataiku empowers its employees to unleash their extraordinary power and generate the AI that powers their organizations into the future.
Dataiku was founded in 2013, and we continue to develop Everyday AI, which has been our vision since then, and continue to grow Dataiku itself.
This has demonstrated its ability. Our blog