Companies spend millions of dollars annually to store terabytes of cloud data—but leverage only a fraction of it for predictive tasks. With Kumo, businesses can unleash the full potential of their enterprise data for faster, simpler, and smarter predictions. Today, in SQL, you already query the past; with Kumo, you can query the future.
Conventional enterprise AI treats every predictive task separately in a silo. However, enterprise data represents a rich, interconnected web of business relationships, interactions, customers, transactions, and more. By leveraging the connectedness of enterprise data, Kumo enables a technical leap-frog in AI.
Kumo’s best-in-class graph learning tech has meaningfully higher accuracy, even with an order of magnitude less training data.
Kumo can train on large-scale databases in minutes (not hours/days).
Automatically go from your raw data at point of storage to making predictions in production robustly over time
All users are supported, from out-of-the-box predictive analytics for analysts to advanced customizations/outputs for sophisticated ML/DS users.
Explain your predictions with the contribution value of every cell of every table in your database.
A single graph instance can support hundreds of real-world predictive tasks that drive business ROI.
Grow your business by predicting Customer LTV, Churn, Affinity, Next Best Action, Pricing, Attribution, and more.
Fortify trust, safety and security by detecting Fraud, Spam, Account Take-Over, Abuse and other toxic behavior.
Curate knowledge with entity resolution, missing value imputation, anomaly detection, search/information retrieval, and more.
Enable power users by exposing embeddings, nearest neighbors, and a wide variety of advanced customization options.
Representation learning led to AI breakthroughs in image and language processing -- but, to date, representation learning methods have been largely limited in applicability to images (matrices) and text (sequences).
Graph Neural Networks (GNNs) generalize representation learning to graph data. GNNs are the only deep learning architecture that is not fixed but adapts to the shape and structure of the underlying data.
Kumo brings GNNs to enterprise data in a way that is both scalable and easy to use. Kumo ingests raw enterprise data and learns to represent it automatically and optimally in graph form, delivering better accuracy even with less training data.
Our mission is to bring the most powerful graph learning approaches, proven in the research world, to all enterprise data—through simple and elegant tooling. We are passionate about the possibility and promise of graph learning. We believe we can scale enterprise prediction by orders of magnitude, while also dramatically widening the circle of people who can effectively use predictive AI.
We are a venture-backed, SaaS AI start-up with a roster of brilliant investors and advisors. Founded by several pre-eminent AI executives from companies like Pinterest, Airbnb, and LinkedIn, we have a multi-year head start in developing Kumo’s technology through the Stanford/Dortmund University research labs and the PyG team. We are hiring and growing quickly.
We believe enterprise data is inherently highly connected—and that insight about connectedness powers our AI tech. We also believe that the best teams are highly connected. We are all nodes in the network; we are better together.
We are optimistic by nature, and in particular about the massive potential of AI and graph learning. The possibility and promise of GNNs, rendered easy and usable for any enterprise to use, motivates our work.
We have seen firsthand the incredible effort to implement a single, production-quality predictive model. With Kumo, the team aims to solve this problem and make graph learning easy to use, so any business (and person and user) can leverage the power of graph-based AI in minutes.
Co-founder and CEO
Co-founder and Chief Scientist
Co-founder and VP of Engineering
VP of Product
Pinterest, Chief Architect
Chief Information Security Officer, Confluent
President & COO, Opentrons
MARCH 15, 2022
Kumo Emerges From Stealth to Launch Next Generation Predictive AI for Businesses; Raises $18.5 Million in Series A Funding led by SequoiaRead More
Our tech is based on connectedness—the interconnection of enterprise data—and so is our team. Like nodes in a (neural) network, we are better together. Join us!
We are building a highly collaborative team to enable next-generation AI. We are deeply optimistic by nature, and passionate about the massive potential of graph learning, implemented scalably across distributed systems. We make graph learning easy to use—so any business (and person and user) can leverage the power of graph-based AI. If this sounds exciting, send us an email at email@example.com.
The strength of our network is based on the well-being of all of our team mates. Our employment philosophy is human-centered, putting well-being and care first.
Gold Plan Health Insurance
Dental & Vision Coverage
Flexible Spending Accounts (FSAs) for Health and Dependent Care
Unlimited discretionary time off
Full-week shutdown at end-of-year
Fully remote and hybrid work as an option
South Bay office, 0.4 miles from Mountain View Caltrain stop