Using Machine Learning to Manage User Access

Due to our commitment to technology, we use many different applications at Capital One. In addition to the plethora of open source tools, we extensively use AWS, on-premises servers, data management and visualization tools, and more. While the seemingly unlimited technology options allow us to create amazing products for our customers, it begins to become difficult to manage access rights to these applications and servers. We have used the Louvain method for community detection to cluster Capital One associates based on their actual work habits and roles. Instead of relying solely on the organizational chart, this leads to a far more accurate and realistic representation of the company. Once these clusters have been determined, we can more accurately detect which accesses an associate needs to be productive. In order to communicate these access recommendations, we've built an extensive dashboard utilizing React for the front-end, Elasticsearch for the data layer, and AWS for the infrastructure resources.

Amanda is a Software Engineer at Capital One with three years of experience in the tech industry. During this time, Amanda has worked in many technical roles, which primarily focused on cloud infrastructure and front-end development. To add to her cloud cred, she has received her AWS Solutions Architect and Developer Certifications, both at the Associate level, and has two publicly available Alexa skills. On the front-end side of things, she not only loves creating modern user interfaces in React, but she's also been titled the local graphic designer by her team members, making multiple logos to represent thousands of people within Capital One.