Follow Datanami:
November 27, 2023

Kumo.AI Unveils New Predictive AI Platform With SQL-Like Querying


Over the past several years, technology has rapidly changed what businesses can do. Artificial intelligence is currently at the center of conversations around the globe. While generative AI creates novel content, predictive models focus on forecasting future trends based on current data. Compared to generative AI, predictive AI can handle more complex and enterprise-specific tasks, but it often lacks adaptability and has limited scope outside its domain. 

Kumo.AI, a predictive AI company, plans to tackle this challenge. It recently announced the general availability of a new predictive AI platform that features SQL-like predictive querying language. This platform will enable enterprises to use their data for the rapid creation and deployment of state-of-the-art AI models. 

The new predictive AI platform can be used to build multiple task-specific AI models in a single day. Kumo.AI is now offering a free trial for enterprises to test its capabilities on their private data.

Founded in 2021, Kumo.AI is a SaaS startup backed by Sequoia Capital. The startup was founded by severe pre-eminent AI executives from companies like Airbnb, Pinterest, and Linkedin. The diverse experience of its founders and the collaboration with Stanford/Dortmund University research labs has allowed Kumo to get a multi-year head start in developing its AI technology. 

Kumo’s predictive AI platform uses graph neural network (GNN) technology, empowering data scientists, analysts, developers, and business owners to create and implement highly accurate predictions in production. 

According to the Co-founder and CEO of Kumo.AI, Vanja Josifovsk, the primary goal of the startup is to empower every data science team so they can address customer-focused use cases dramatically faster without compromising on accuracy. 

Josifovsk also said, “Kumo.AI is redefining predictive AI by leveraging Graph Neural Networks, a technology pioneered by my Co-founder, Prof. Jure Leskovec at Stanford. We have built the most scalable and efficient graph learning platform, capable of generating predictions for hundreds of millions of entities, such as providing product recommendations for millions of users every day.”

Kumo.AI’s new platform uses deep representation learning technology to streamline the most complex and time-consuming steps in the machine learning process. It achieves this by eliminating training data generation, feature engineering, feature pipelines, and feature stores. Kumo.AI is the first platform to apply deep learning representation to enterprise data in data warehouses. 

At the Snowflake Summit in June this year, Kumo.AI announced the integration of its deep learning capabilities directly into the Snowflake data cloud through Snowpark Container Services to maximize the data potential by using cutting-edge tools to eliminate the need for data movement. 

Kumo is transforming the future of business applications, making predictive AI accessible and practical for enterprises of all sizes. With the launch of the new platform, Kumo is one step closer to its goal of revolutionizing the way enterprises leverage predictive AI.

Related Items 

Will Mass Adoption of GenAI Lift Traditional AI?

Pecan AI Opens Platform and Fast Predictive Modeling for All Data Professionals

The Mystery of Predictions in AIOps