RiseLab UC Berkeley logo

Capital One Joins Founding Sponsors of RISELab at UC Berkeley

Capital One DevExchange
Capital One Tech
4 min readJan 23, 2017

--

RISELab’s mission is to enable machines to rapidly take intelligent actions based on real-time data and context from the world around them.

By Stephen Romanoff

As a fundamentally data-driven company, Capital One has leveraged insights from data for decades to offer our customers financial products that meet their diverse needs. Now, we are entering a new era in data and computing premised upon the foundation of big data technologies and cloud scale, and characterized by big/fast data, automated intelligent decisions, and holistic security.

The convergence of these technologies could unlock myriad new opportunities to harness data and context in real time to help our customers with their financial lives. This will be critical as we increasingly move toward a future where banking is more real-time, digital and better able to anticipate customer needs.

It is this focus that spurred Capital One to join the founding sponsors of a new initiative at the University of California Berkeley’s Electrical Engineering and Computer Science Department, the RISELab. Over the next five years, the RISELab will focus on developing a secure real-time decision platform, including tools and algorithms, to enable fast, automated decisions on live data. The RISELab is built on the model of Berkeley’s remarkably successful AMPLab, which developed foundational open source big data capabilities including Apache Spark, Apache Mesos, and Alluxio.

The mission of RISELab — to enable better, more secure real-time decisions with a high-degree of automation — and its focus on metadata and contextual data as foundational to enabling other components; advanced computing frameworks that enable speed and scale; material advancements in machine learning on live data; and robust end-to-end security — could lead to breakthrough customer experiences across industries.

Over January 11–13, we joined UC Berkeley RISELab faculty and students -along with other founding sponsors Google, Amazon Web Services, Microsoft, GE, Intel, IBM, Ericsson, VMWare, Huawei, and ANT Financial - to kick off the RISELab initiative.

In presentations and discussion, we learned more about the strategy, objectives, and technical details of the lab. Below are a few highlights of the exciting projects the lab will undertake:

· Ground is a data context service which enables managing all data about data in a central, agnostic manner. The project’s perspective to add temporal and behavioral context to traditional metadata, present an opportunity to better leverage this contextual data in a complementary way to how people use data. As a data-driven company, Ground is particularly valuable to Capital One and we have already begun to collaborate with the RISELab on this initiative. James Dalton, a Senior Data Engineer on our team, is working together with UC Berkeley researchers in the development of Ground, which is a foundational component of the Secure Real-time Decision Stack.

· Clipper is a modular framework for optimizing and deploying models across machine learning frameworks from user applications. As Capital One’s use of machine learning increases across diverse applications, frameworks like Clipper will enable us to simplify model deployment and improve our performance and accuracy.

· Capital One is leveraging graph structures to better identify and find new relationships between disparate data elements and to aide in use cases that better anticipate customer needs. Tegra enables efficient analysis of graph-structured data as the graph evolves. This temporal context provides the ability to query across changes in the underlying graph and allows for us to explore not just relationships at a point in time, but what they were and how they evolved. This ability to look back at the evolution of a relationship is powerful.

· As Capital One continues its adoption of cloud technologies, we are always looking for new ways to enable our teams and users to choose the best VM for their specific workload. Paris is a modeling framework for estimating workload performance across VMs and even across cloud providers. This capability can better enable us to migrate to the cloud more efficiently and cost-effectively.

In addition to the promising technology research and development of the RISELab, Capital One’s sponsorship also furthers our commitment to open source software development and the Bay Area technology community. By partnering with academic and industry partners, our aim is to collaboratively solve pressing technology challenges that will lead to breakthrough customer experiences in our industry and beyond.

The work of the RISELab is just beginning. I encourage you to check out the projects and people of the lab and to check back on this blog for updates.

For more on APIs, open source, community events, and developer culture at Capital One, visit DevExchange, our one-stop developer portal. https://developer.capitalone.com/

--

--

Capital One DevExchange
Capital One Tech

Opinions and thoughts on API and Open Source development at Capital One.