Inc

My journey into data science didn’t start in computer science or statistics—it began in psychology. I was captivated by cognitive science and neuroscience, especially the ways we can use computational models to better understand human behavior. As that curiosity grew, I realized I needed a deeper mathematical foundation to answer the kinds of questions I was asking. That realization led me to pursue a second degree in applied mathematics, and later a master’s in mathematics and statistics at Georgetown University, where I focused on data mining, stochastic processes, and machine learning.

After graduate school, I joined IBM, where I worked on applying machine learning to complex, high-impact challenges in government and healthcare. My projects included fraud detection and risk modeling, areas where data-driven decision-making can have real-world consequences. That experience taught me a lot about the messy realities of deploying machine learning in large, multifaceted organizations.

My transition into media and entertainment came when I joined Audacy, a national leader in radio and podcasting. What drew me in was the chance to apply advanced data science to a dynamic industry with rich, real-time data. From personalization and audio-transcription models to forecasting, retrieval-augmented generation (RAG), and large-scale experimentation, the possibilities felt endless. I’ve since grown into the role of Director of Data Science, where I lead a cross-functional team of data scientists and ML engineers focused on driving innovation aligned with Audacy’s broader business strategy.

Aligning Data Science with Strategy and ROI

To ensure every data science initiative delivers measurable value, I rely on a concept borrowed from education: scaffolding. As scaffolding supports learners through complex material, structured frameworks help data science teams navigate uncertainty and stay focused on strategic priorities.

At Audacy, we operate in Scrum sprints, allowing our data science work to align tightly with engineering and product. This shared cadence strengthens communication and helps us integrate models smoothly into larger platforms. We follow the data science lifecycle and MLOps frameworks rigorously, from initial business understanding through data exploration, modeling, deployment, and monitoring.

 ​What makes our experimentation framework successful is the connection between technical rigor and business context. It’s not just about proving something works—it’s about understanding why it works, who it benefits, and how it fits into our company’s larger goals 

One key focus is data quality. Even the most sophisticated models can’t perform well on messy or inconsistent data, so we invest early in ensuring our data is reliable and accessible. Before any modeling begins, we align with stakeholders on success metrics and clarify the business decisions we aim to support. For more exploratory efforts, we use flexible frameworks like OSEMN or Kanban. But as we move toward production, structure, governance, and automation take center stage. This combination of flexibility and rigor helps us deliver impact transparently and at scale.

Driving Change through Communication and Trust

One of the biggest hurdles I’ve faced is translating complex insights into something that resonates with non-technical stakeholders. With a background in applied math, I love diving into the details of a model, but I’ve learned that technical precision alone doesn’t drive decisions. What matters more is answering the question: So what?

If a model improves accuracy by five percent, what does that mean for the business? Will it increase revenue? Enhance customer satisfaction? Improve operational efficiency? When we can clearly articulate that impact, it becomes much easier for leaders to take action.

Building trust across departments, from product and engineering to editorial and executive leadership, requires delivering timely, actionable, and directly connected insights to business outcomes. At Audacy, we also foster a culture of analytics through workshops, transparency, and shared success metrics. These practices help make data science not just a support function but a strategic driver.

Establishing a Culture of Experimentation

When I joined Audacy, one of my first goals was to establish a formal A/B testing program. At the time, product decisions were not experimentally driven. I partnered with the Product team to build a framework that supported structured, data-driven experimentation.

Good experimentation starts with clear hypotheses. For instance, if we’re testing a new app layout, we don’t just ask whether users like it—we ask whether it improves listening hours or enhances engagement, and we define specific metrics to measure those outcomes. We also monitor counter-metrics to catch any unintended consequences.

A/B testing is especially powerful when applied to machine learning. Before launching a recommendation system, for example, we test different models in controlled environments, evaluating metrics like click-through rates and time on the platform. We typically start small, testing on limited user groups, to manage risk and gather insights before scaling up.

What makes our experimentation framework successful is the connection between technical rigor and business context. It’s not just about proving something works—it’s about understanding why it works, who it benefits, and how it fits into our company’s larger goals.

Leading with Purpose in a Fast-Paced Industry

If I could offer one piece of advice to data science professionals looking to lead in consumer- focused industries like media and entertainment, it would be this: invest in your people.

At Audacy, I’m fortunate to lead a team of incredibly talented data scientists and machine learning engineers, each bringing unique strengths and perspectives. Regular knowledge- sharing sessions are among the best ways to harness that diversity. Whether exploring new modeling techniques, reviewing academic research, or debriefing a deployment, these sessions promote curiosity, collaboration, and collective growth.

Leadership, especially in a fast-moving field like ours, requires active engagement. It means understanding what your team is building and what they need to succeed—whether that’s resolving a technical blocker, securing resources, or connecting their work to business strategy. I embrace the Servant-Leader model: my role is to empower others to do their best work, challenge each other constructively, and continue evolving as professionals.