Capital One Banks on Snowflake for Data Science

Capital One's mission is to change banking for good when we envision a future for banking. We don't start with what other banks are doing. We really take a look at how technology is changing people's lives. 

I am Salim Syed. I'm the engineering lead. 

My name is Nagender-Gurram, Director, Data Engineering. I'm 16 years at Capital One and I have seen multiple transformations. The biggest step that we did was going into the cloud. 

Things are changing really fast and we needed to adapt. We had to be agile. Going to the cloud was the first step. Other financial companies are going in the cloud, but they're tiptoeing in. We're the first bank to move all public cloud. We had to figure out what are the best practices with unlimited compute and storage. How do you manage governance in this new world? 

We found Snowflake to be a revolutionary concept.

Snowflake gave us the separation of compute and storage, which helped our data engineering teams focus more time on innovation than worrying about managing or provisioning the infrastructure. 

It's really the amount of data that's coming in now and at the pace, the speed that is coming in. Do you have the infrastructure to get the insights quickly for your customers? Now, if there is a crash on a cluster Snowflake automatically restarts and picks up where it left off. And that's huge for our business continuity. 

With Snowflake, our data scientist community can create machine learning algorithms and train the models so that we can sell a lot of digital products like Eno, which is our Capital One Intelligent Assistant. When we first selected Snowflake, we realized we had to start building self-service tools that we could just give to our lines of business so they could manage their own onboarding of Snowflake resources. 

The point was to make it easy, transparent, hide the governance complexity, and that allowed us to move much faster.

We use Amazon Web Service as our primary cloud provider, and Snowflake integrates really well with it. The way we move data was extremely fast because of the integration that they've already made. We have thousands of analysts running millions of queries a day, and sometimes we're having a few concurrent queries. They're getting predictable performance. Their queries are running at the same speed, no matter who else is running on the platform. 

We are excited about a lot of new Snowflake features—things like Snow Park, where we can better implement our machine learning models. Data marketplaces is another place where Capital One is exploring. We're running multiple POCs, we're already using COVID-19 data, and it's really helping us make better financial decisions around where the hotspots are and how to react quickly.

It's completely revolutionizing the data industry. We have a great partnership. Snowflake is always open to take the feedback. I think Snowflake has been hugely powerful. It saves a lot of time, a lot of money. Speed to market increases can scale to the demand. 

Simple answer is Snowflake met our use case better than any other technology out there.