When most people think about jobs in finance, they picture investment banking, accounting, or maybe consulting. Fraud analytics is usually not the first thing to come to mind. But after talking to Justin De La Rosa, and alumni from our graduate program, it is clear that this is one of the most interesting and growing corners of the financial tech world.
Justin De La Rosa currently works as a Fraud Analytics Analyst at Live Oak Bank and explained that fraud analytics sits right at the intersection of data, technology, and real-world problem solving. As more banking happens online, financial institutions are relying heavily on data and machine learning to spot suspicious activity before it turns into real losses.
Justin shared that a lot of their work involves building analytics frameworks, setting up data pipelines, and pulling information from different sources to get a full picture of what is happening. Across the industry, banks use a mix of internal data and third-party platforms that specialize in fraud detection. These tools constantly scan for unusual patterns, things like logins from unexpected countries, strange timing, or sudden changes in behavior, and then flag them for further review.
Machine learning plays a big role in making all of this possible. According to the Justin, many banks use vendor platforms that already have machine learning models built in, rather than creating everything from scratch. Models like gradient-boosted trees are especially popular because they work well and are easier to explain than more complex neural networks. In finance, being able to explain why a model flagged something is just as important as the prediction itself.
What made the conversation especially interesting was hearing how transferable analytics skills really are. Before moving into fraud analytics, Justin worked at Wells Fargo in a marketing analytics role. At the bank, they focused on predicting customer growth, estimating the return on ad campaigns, and helping teams decide where to spend money. Even though the business goals were totally different, the core skills, working with messy data, building models under uncertainty, and communicating results, translated directly into fraud and risk work.
One thing Justin emphasized over and over was communication. Being good at analytics is not just about coding or modeling, but explaining what the results mean and why anyone should care. Whether talking to stakeholders the ability to adjust your message to your audience makes a huge difference. According to Justin, this is easily the most transferable skill across roles, companies, and industries.
For students who want to get into fraud analytics or Data Analytics in general, the advice was refreshingly easy to implement. Do projects. Get comfortable with the entire analytics lifecycle. Be ready to talk about your work in a clear, confident way. Just as important, be personable and curious, people remember how easy you are to work with, not just what tools you know.
Looking ahead, Justin De La Rosa sees fraud analytics continuing to grow as AI and machine learning become more deeply embedded in financial systems. As fraud becomes more sophisticated, banks will need analysts who understand both the technical side and the bigger picture. For students who like data, technology, and solving real problems, fraud analytics might be a career path worth paying attention to.
Columnist: Lex Chaffee