Melissa Lee: From stay-at-home mom to data scientist detecting bad guys
Melissa and other data scientists do this on our network on a much larger scale. They learn what normal traffic looks like and they pick out the things that don't belong there.
"We do that using singular value decomposition," she said. "That's just a sneaky math trick to identify the patterns in the data."
Hackers and malicious actors, though, make the job difficult. They would force Melissa's model to perceive bad behavior as normal. But Melissa got wise and learned, too, employing a different analytic technique, called the robust method.
"Basically, we let math separate the noise before we build our model, so we catch the bad activity," Melissa said. "Now think of it as a shopping mall with a lot of people walking around. You only want to see the actual mall. You can use a robust method to separate the people from the surroundings to get a more accurate picture."
Super math skills to the rescue
Every day is a new challenge for our data scientists. Lately, Melissa has been working on fraud and risk analysis. She works to protect our consumers from a possible account-takeover situation when they log into websites.
"You feel a little bit like a superhero," Melissa said. "It's almost a gotcha! gratification when you detect something that someone is trying to hide from you. The idea of knowing you're getting ahead of the bad guy just a little bit is really nice."