Kumo

As Sridhar Ramaswamy noted on Wednesday’s earnings call, Kumo is succeeding alongside Snowflake thanks to “a great product, a sales team that executes at scale, and the ability to introduce new initiatives…” Companies use Kumo to train predictive and embedding models directly on their data in Snowflake, without time-intensive feature engineering. Those models already power decisions inside business-critical applications that reach hundreds of millions of users globally. You can try Kumo for free as a Snowflake Native App: https://lnkd.in/dp6d-nnZ

The AI x Relational workshop in SF on March 12 will feature Ben Shahshahani, Chief AI Officer at Cleveland Clinic, discussing real-world applications of AI models for relational data. Register now for the in-person event. Space is strictly limited: https://lu.ma/6ax743pt

Combining LLMs with relational AI models from Kumo improved recommendation accuracy by 15x (MAP@12) vs. just using the LLM. https://hubs.ly/Q037Dq560

I am excited to be doing a keynote and panel with Joaquin Quiñonero Candela from OpenAI and Sakshi Jain (LinkedIn) and Grace Tang (LinkedIn) in March. Registration information below. We will be talking about building AI systems in a Responsible way.

Tomorrow: Live workshop on Graph Transformers, featuring experts from NVIDIA - Rishi Puri, Kumo - Matthias Fey, and Stanford University - Vijay Prakash Dwivedi. Last chance to register:  https://lnkd.in/dTSWhfwn

RelBench is the first public benchmark to systematically study predictive tasks over relational databases using Graph Transformers. It provides datasets and tasks across a variety of domains and scales, from small relational structures to large, complex schemas. Our goal was to create a foundational infrastructure that researchers and practitioners can use to evaluate, compare, and improve ML approaches on relational data. https://hubs.ly/Q037D8h30

Yieldmo trained Kumo AI models on Snowflake to drive real value (+15% ad engagement) in a fraction of the time it'd take with traditional ML. See how they did it, from the Principal Data Scientist at Yieldmo: https://lnkd.in/dThmvdSK

Announcing AI x Relational: An in-person workshop and meetup by Kumo featuring Jure Leskovec Train your first graph-transformer-based AI model directly on your data warehouse without feature engineering. Join us in San Francisco on March 12 to explore techniques that are transforming predictive modeling and AI applications using breakthrough model architectures and the relational data you already have. - Technical deep-dives on Graph Transformer architectures and AI models for relational data - Production implementation case studies from teams handling millions of transactions - Hands-on workshop with sample datasets - Networking with AI practitioners who've deployed these systems at scale Limited to 50 participants to ensure meaningful technical discussions. March 12, 2025 | 4:30-8:00 PM | Amazon SFO28, San Francisco Register: https://lnkd.in/dazPjEki

Data scientists and engineers can now build predictive and embedding models directly on company data in Snowflake. Eliminate feature engineering and reduce model development time by up to 95% compared to traditional methods. Try it free as a Snowflake Native App: https://hubs.ly/Q037C-xw0

Switching from traditional predictive models to graph-transformer based predictive models doesn’t mean losing touch with the model. Every AI model you train on Kumo comes with explainability to help you trust, understand, explain, and iterate your models. Read the docs: https://hubs.ly/Q037D0yt0 Try it: https://hubs.ly/Q037CXkg0