Graph Learning at the Scale of Modern Data Warehouses

Data warehouses have become a staple for enterprises, providing a wealth of information that can be harnessed to improve decision-making through the use of machine learning (ML). The data stored in these warehouses is typically relational and can be viewed as a graph of entities and relationships. Graph learning with graph neural networks (GNNs) offers a natural and effective way to apply ML to this type of data. However, deploying GNNs at scale presents several challenges, including transforming the data for graph learning, scaling the graph learning framework, and performing predictions in a reasonable amount of time. This presentation will outline our comprehensive approach to addressing these challenges and show how we built an efficient and scalable end-to-end system for graph learning in data warehouses.


Subramanya Dulloor

Founding Engineer

Subramanya Dulloor is a computer scientist and engineer with extensive experience in distributed systems and systems for ML. As a founding engineer at, he is focused on making predictive analytics on data warehouses plug-and-play easy. Before joining Kumo, Dulloor worked on distributed systems and database internals at various startups. He also spent several years at Intel Labs conducting research in memory management in operating systems and VMMs, as well as in distributed high-performance systems for machine learning and computer vision.

Dulloor holds a PhD in computer science from Georgia Tech and has over a dozen publications in top-tier conferences and patents to his name. 

Read more
Find Subramanya Dulloor at:


Wednesday Mar 29 / 02:55PM BST ( 50 minutes )


Mountbatten (6th Fl.)


Graphs Data warehouse Machine Learning


From the same track

Session Machine Learning

Strategy & Principles to Scale and Evolve MLOps @DoorDash

Wednesday Mar 29 / 10:35AM BST

MLOps has become a major enabler to successfully operationalize ML applications and for ML practitioners to realize the power of ML to bring impact to business.  The journey to implementing MLOps will be unique to each company.

Speaker image - Hien Luu

Hien Luu

Sr. Engineering Manager @DoorDash & Author of Beginning Apache Spark 3, Speaker and Conference Committee Chair

Session AI

Responsible AI: From Principle to Practice!

Wednesday Mar 29 / 01:40PM BST

Enabling responsible development of artificial intelligent technologies is one of the major challenges we face as the field moves from research to practice. Researchers and practitioners from different disciplines have highlighted the ethical and legal challenges posed by the use of machine learn

Speaker image - Mehrnoosh Sameki

Mehrnoosh Sameki

Principal PM Manager @Microsoft

Session Digital Twins

Cognitive Digital Twins: A New Era of Intelligent Automation

Wednesday Mar 29 / 11:50AM BST

Traditionally, Digital Twins have been helping businesses make data-driven decisions, increase efficiency, and improve the overall performance of their physical assets.

Speaker image - Yannis Georgas

Yannis Georgas

Intelligent Industry Lead @Capgemini

Session python

Simplifying Real-Time ML Pipelines with Quix Streams: An Open Source Python Library for ML Engineers

Wednesday Mar 29 / 04:10PM BST

As data volume and velocity continue to increase, the need for real-time machine learning (ML) is becoming more pressing. However, building real-time ML pipelines can be complex and time-consuming, requiring expertise in both ML and streaming application development.

Speaker image - Tomáš Neubauer

Tomáš Neubauer

CTO & Co-Founder @Quix