Time’s almost up! There’s only one week left to request an invite to The AI Impact Tour on June 5th. Don’t miss out on this incredible opportunity to explore various methods for auditing AI models. Find out how you can attend here.
Every year, Snowflake hosts the Data Cloud Summit to share how it is expanding its core platform to help teams make the most of their data assets. The year 2024 was no different, but the focus was more centered on one key aspect — enabling teams to build powerful AI applications with their data assets.
The Sridhar Ramaswamy-led company announced several capabilities and partnerships focused on AI innovation, while also releasing some long-awaited features to strengthen the data foundations of the platform, including the general availability of Iceberg Tables and an internal marketplace.
Below is a rundown of all major announcements:
Iceberg Tables in GA and the new Polaris Catalog
First, the company shared that it is making Iceberg Tables generally available, unlocking full storage interoperability for enterprises. Iceberg Tables work like Snowflake native tables but store the table metadata in Apache Iceberg format in the customer-supplied storage. This takes Snowflake’s ease of use, performance, governance and collaboration to their Iceberg data stored externally.
June 5th: The AI Audit in NYC
Join us next week in NYC to engage with top executive leaders, delving into strategies for auditing AI models to ensure optimal performance and accuracy across your organization. Secure your attendance for this exclusive invite-only event.
The company further reinforced its commitment to Iceberg with the launch of Polaris, a vendor-neutral, open data catalog implementation for indexing and organizing data conforming to the open table format.
Available in both self and Snowflake-hosted options, the catalog will be open-sourced over the next 90 days and interoperate with other query engines enterprises would like to use to drive value from their data assets, including Apache Flink, Apache Spark, Dremio, Python, Trino and others.
Cortex AI and Snowflake ML upgrade
Snowflake’s Cortex AI, the fully managed service to build large language model (LLM) applications, got a major upgrade with an AI & ML Studio, a no-code interface to get started with the testing and evaluation bits of the application development workflow. Cortex is also getting new Analyst and Search offerings, with the former helping with the development of LLM chatbots that help users ask questions on their business data, while the latter helping with retrieval augmented generation (RAG) and enterprise search apps that use documents and other text-based datasets through enterprise-grade hybrid search.
The company also announced Cortex Guard as a safety net that ensures these chatbots do not produce harmful content in their responses as well as new MLOps capabilities in Snowflake ML.
Snowflake Horizon’s new marketplace
Snowflake launched Horizon to provide organizations with a built-in set of compliance, security, privacy, interoperability and access capabilities to discover and govern data, apps and models. Now, the suite is being enhanced with a new internal marketplace that allows users to curate and publish these data products specifically for teams within their organization to discover and use. It also includes control options to help users limit who within an organization could see or access their content. The offering will also support the sharing of AI models, Iceberg Tables and Dynamic Tables, the company added.
Additionally, Snowflake said it is introducing new AI-powered object descriptions that will automatically generate relevant context and comments for tables and views, enabling better data discovery and curation.
Snowflake Trail for observability
In addition to data and AI-centered improvements, Snowflake focused on the observability side with the launch of Trail, an offering that provides visibility into data quality, pipelines and applications. Trail provides built-in telemetry signals for Snowpark and Snowpark Container Services, enabling users to easily diagnose and debug errors using metrics, logs, and distributed tracing — without sitting up agents or transferring data manually.
Trail uses OpenTelemetry standards, which means it can also be integrated with other observability and alert platforms, including Datadog, Grafana, Metaplane, PagerDuty and Slack.
Stronger data residency
Snowflake also said it is expanding its AI Data Cloud footprint to some highly regulated and sovereign markets globally. The company did not share the names of these regions but confirmed this includes a new EU-only data boundary that will keep all customer data, alongside relevant service and usage data, within regional borders to ensure stronger data residency and help customers meet regulatory requirements.
The company will also offer a separate environment to Department of Defense customers with stronger security and a networking integration with Boundary Cloud Access Point (BCAP).
The Nvidia partnership
Finally, Snowflake announced a partnership with Nvidia to adopt the Nvidia AI Enterprise software and integrate NeMo Retriever microservices into Cortex AI. The engagement will enable organizations to connect custom models to diverse business data and get highly accurate responses.
Beyond, the company said its open LLM Arctic supports Nvidia TensorRT-LLM software, promising highly optimized performance. The model is also now available as an NVIDIA NIM inference microservice, allowing more developers to access it.
Snowflake Data Cloud Summit runs from June 3 to June 6, 2024.
Source link lol
