New release enables organizations to detect and prevent unexpected AI behaviors at-scale across AI assets, environments and teams
DataRobot, the enterprise AI platform, launched new AI observability functionality with real-time intervention for generative AI solutions, available across all environments including cloud, on-premise and hybrid. This latest release provides AI leaders and teams with the tools to confidently build enterprise-grade applications, manage risk and deliver business results.
“Lack of visibility and risk are significant obstacles to reaching real business value from AI,” said Venky Veeraraghavan, Chief Product Officer, DataRobot. “We’re revolutionizing AI observability with real-time intervention across diverse AI assets and environments, so leaders can safeguard projects, up-level oversight and empower teams.”
Today’s announcement brings AI observability for any AI asset and environment into the DataRobot AI Platform to deliver:
- Cross-Environment AI Observability: Gain full oversight across environments and reduce risk across your entire AI landscape with unified governance for all predictive and generative AI assets.
- Real-Time Generative AI Intervention and Moderation: Build a multilayered defense to safeguard AI applications with customized build, intervention and moderation workflows, leveraging a rich library of pre-built and configurable guards to ensure accuracy and prevent issues like prompt injections and toxicity, detect personally identifiable information (PII) and mitigate hallucinations.
- Generative AI Alerts and Diagnostics: Gain control and flexibility with customizable alert and notification policies, visually troubleshoot problems and traceback answers, and set robust multi-language diagnostics with insights for data quality checks, topic drift and more.
“Today’s AI leaders are seeking to implement generative AI accurately, safely, and at scale,” said Amanda Saunders, Director, Enterprise Generative AI Product Marketing, NVIDIA. “DataRobot’s integrations with NVIDIA software provide enterprises with purpose-built generative AI solutions to speed successful deployments.”
This new release also introduces best-in-class evaluation, testing and open source LLM support capabilities:
- Enterprise-Grade Open Source LLM Hosting: Leverage any open source foundational model including LLaMa, Hugging Face, Falcon and Mistral with DataRobot’s built-in LLM security and resources, complementing recent integrations with NVIDIA NIM inference microservices and NVIDIA NeMo Guardrails software to accelerate AI deployments for enterprises.
- LLM Evaluations, Testing and Metrics: Enhance application quality, assess LLM performance and automate testing with groundbreaking out-of-the-box synthetic test data creation, evaluation metrics and quality benchmarks.
- Advanced RAG Experimentation: Evaluate different embedding methods, chunking strategies, and vector databases to assess and identify the best RAG strategy for each use case.
”DataRobot is providing the requisite tools an enterprise needs to get started with generative AI quickly and with confidence,” said Ritu Jyoti, Group Vice President, Worldwide Artificial Intelligence and Automation Research Practice Global AI Research Lead, IDC. “AI leaders need to be prepared for a new normal where AI exists in multiple environments and locations. The work DataRobot is doing with AI observability sets the foundation for AI leaders and teams to thrive.”
This release is the latest milestone in DataRobot’s significant investment in generative AI for the enterprise. In the last 12 months, DataRobot introduced industry-first generative AI functionality, launched the DataRobot Generative AI Catalyst Program to jumpstart high-priority use cases, and announced new and expanded collaborations with global leaders like NVIDIA and Google Cloud to supercharge AI solutions with world-class performance and security.
Source: BusinessWire
