Platform pitched as reducing AI development time by 60% and handling AI workloads 19x faster; allowing enterprises to create AI that increases in accuracy as customers engage.
DataStax says the DataStax AI Platform, built with NVIDIA, reduces AI development time by 60%.
These platform integrates the DataStax AI platform with NVIDIA AI Enterprise software and is pitched as making it easier for enterprises to build AI applications that leverage companies’ enterprise data and context.
The platform is also pitched as making it easier for enterprises to hone their models so they self-learn and get more accurate with customer use.
Kari Briski, NVIDIA, Vice President AI Software, said: “Enterprises are harnessing AI to drive digital transformation across industries. The DataStax AI Platform, built with NVIDIA AI, enables companies to create AI-ready databases and rapidly deploy tailored AI applications, unlocking new levels of customer value.”
The Platform offers enterprises a complete AI solution for all parts of the AI development and production lifecycle – from data ingestion and retrieval to application development to deployment and ongoing AI training.
This is said to solve two urgent problems:
AI Application Development: There are new tools necessary for AI application development, and those tools are mostly built for individual developers. However, they tend to break when it comes to the team and organizational workflows necessary in larger organizations. Right now, most AI projects are failing or are behind due to tool complexity. Having a unified platform greatly simplifies the variability in the tech stack.
Enables More Accurate AI: Companies need their AI applications to generate output based on their enterprise data, relevant to their business domain, and to learn from their customers’ interactions. NVIDIA NeMo Customizer and NeMo Evaluator simplify training or fine-tuning LLMs, SLMs, embedding models, and reranking models while DataStax’s AI application development platform gives developers the dynamic control of search and retrieval that is necessary to tailor GenAI to individual customers.
Key pieces in the Platform include:
DataStax Langflow platform, including an application development environment that simplifies creating and understanding complex logic flows via an intuitive visual interface.
NVIDIA AI Enterprise components, including:
NeMo Retriever: Enable organizations to seamlessly connect custom models to diverse business data and deliver highly accurate responses. This is crucial for companies’ AI applications to have the context of their enterprise systems and data in order to provide the most relevant results.
Multimodal PDF Data Extraction: A blueprint that enables ingestion of unstructured and complex enterprise data sources such as PDFs to be ready for use for AI applications using retrieval-augmented generation (RAG).
NeMo Curator: Data-curation tool that helps developers create large, high-quality datasets for pretraining or fine-tuning LLMs or embedding models.
NeMo Customizer: High-performance, scalable microservice that simplifies fine-tuning and alignment of LLMs for domain-specific use cases. This reduces dependencies on off-the-shelf, broad LLMs and enables increased accuracy at lower cost.
NeMo Evaluator: Automates evaluation of customer generative AI models and LLMs for accuracy.
NeMo Guardrails: Easily add programmable guardrails to LLM-based conversational applications.
NIM Agent Blueprints: A catalog of pretrained, customizable AI workflows that equip developers with a full suite of software for building and deploying generative AI applications for canonical use cases, such as customer service avatars, retrieval-augmented generation, and drug discovery virtual screening.
DataStax Data Management: AI requires the most diverse data needs of any enterprise application. Companies have found that they want an integrated data solution versus the complexity of bolting on different solutions for vector or knowledge graph capabilities. DataStax delivers industry-leading vector search, flexible hybrid search, knowledge graph and graph RAG, real-time AI analytics, streaming, pub/sub, and a linearly scalable NoSQL store. Available in the cloud (DataStax Astra), or cloud-native self-managed software (DataStax Hyper-Converged Database).
The Platform is for both cloud and self-managed environments, giving enterprises the flexibility to deploy as they prefer: cloud deployments can leverage their Amazon Web Services (AWS), Microsoft Azure or Google Cloud environments.
Many large enterprises need to run their AI applications in cloud-native self-managed data centers to have full control over all aspects of their technology stack. This is especially true for heavily regulated industries like banks, insurance companies and healthcare companies, which have often had issues with other AI tools that weren’t built for enterprise scale or compliance needs.
Sandeep Varma, PhysicsWallah, Head of AI, said: “PhysicsWallah is democratizing education through GenAI-driven learning experiences for over 20 million students in India. The DataStax AI Platform, built with NVIDIA AI provides a real-time solution for PhysicsWallah to offer personalized, high-quality learning and accessibility at scale. This partnership enables the company to manage a 50x surge in traffic with zero downtime, serving millions of students.”
Chet Kapoor, Chairman & CEO, DataStax, said: “As companies strive to leverage AI, we’re laser-focused on simplifying and accelerating the path to production to unlock innovation at scale. The DataStax AI Platform, built with NVIDIA AI, provides an end-to-end solution that not only reduces cost but unlocks unmatched speed of development — it makes applications smarter and more accurate as customers use it. We’re excited to deliver a platform that will change the trajectory of enterprise AI and redefine customer experiences.”
