Rockset and the Next Era of Search and Analytics for AI

Rockset is on a mission to redefine the data cloud architecture for speed, efficiency and simplicity in an era where real-time AI adoption is growing at a breakneck pace.

“The emergence of generative AI bringing multiple use cases for different companies” - DALL-E

As I wrote in my last blog post The AI Gold Rush For Infrastructure Companies – in this age of Generative AI, we will see the emergence of infrastructure that is optimized for various aspects of AI applications. This creates opportunities for new players to provide “native” solutions for these AI use cases. As a VC, I look for companies that fit our thesis and am excited to announce that we at Icon Ventures are partnering with Rockset to build the future of search and analytics as the world shifts to generative AI.

Built for Familiarity, Not Complexity

As I mentioned in my previous blog, vector embeddings are one of the key ingredients of success in many AI applications. Rather than matching on specific keywords, vector embeddings allow for search of concepts and data that are most similar. Tools like Langchain and Llama-Index make it possible to orchestrate language-based applications while abstracting away a lot of the complexity. Since these models transform text, images, audio and video into vector embeddings, organizations need powerful ways to index these vector embeddings to build modern AI applications.

Rockset is well poised to lead the next generation of search as organizations are rapidly adopting new AI solutions. While it's a full-featured SQL database that organizations already know how to use, Rockset is built for the cloud with compute-compute separation, guaranteeing resource isolation and scalability that is required to index streaming data and search across billions of vectors. Not only that, it's a real-time, all-in-one indexing solution that makes it possible to run search, analytical and vector search queries efficiently in the same database, providing the best of both worlds for any organization.

Strong Roots Made for Growth

Whatnot - the fastest-growing marketplace in the US, is disrupting e-commerce by personalizing live video auctions using AI with Rockset. It is an example of an application that needs to ingest and analyze data in real time. To meet its application requirements, Whatnot needed a fully managed, developer-enhancing platform with real-time ingestion and query speeds, high concurrency and automatic scalability. After thorough diligence, it decided that Rockset was the one that fit its requirements the best.

The Rockset team possesses a deep knowledge and understanding of distributed data systems and indexing engines with their roots in Meta engineering. Dhruba Borthakur created RocksDB, Venkat Venkataramani built and scaled TAO that powers Facebook’s social graph, Tudor Bosman founded the Unicorn project that powers all search at Facebook, as well as building out the infrastructure for Facebook AI Research Lab.

Amongst other things, what impressed us were Rockset’s customer testimonials talking about how simple Rockset made it to build and scale modern data apps, and how Rockset has completely stripped away all the operational and administrative burdens of previous generation solutions such as Elasticsearch. This fits with what we’re seeing in how ML and product engineering teams embrace AI technologies – they want to focus on building AI-powered apps, without the headache of operating burdensome data management systems.

Enterprises like Jetblue, Whatnot and Seesaw use Rockset to build their search and analytical applications at scale. Our investment in Rockset will further bolster its leadership and help take the next generation of search and analytics to new markets.