Ben Berkman joins the show to explain the actual work of a Data Scientist. He specifically works on identity graphs at The Trade Desk. Host Larry Port asks Ben to break down how online advertising auctions happen in a fraction of a second. Ben describes his daily routine, which involves about six hours of coding in Scala and two hours of meetings. He clarifies the distinction between data scientists who build models and the software engineers who build the infrastructure to support them.

ㅤ

Ben also shares how his background in economics and liberal arts helps him ask better questions. He offers an honest look at work-life balance and how AI tools like Claude are changing the way he codes.

ㅤ

Guest Bio

Ben Berkman is a Graphs and Identity Data Scientist at The Trade Desk. He specializes in building data structures for cross-device identity resolution. Before this role, he worked as a Cost Analyst and Data Scientist at Technomics, Inc., where he focused on defense acquisition data. He holds a Master's in Data Science from NYU and an undergraduate degree from Pennsylvania State University.

ㅤ

What We Cover

1. How The Trade Desk facilitates real-time ad auctions for the open internet.
2. The specific breakdown of a data scientist's day: mostly solitary coding with some team collaboration.
3. Differences between data science (creative modeling) and software engineering (plumbing and infrastructure).
4. Why curiosity and communication skills from a liberal arts background are valuable in technical roles.
5. How AI tools are shifting coding workflows from manual typing to agentic oversight.
6. The personality types that thrive in data science: curious problem solvers who enjoy steady work.
7. Realities of work-life balance in a global company with teams in Singapore.

ㅤ

Resources Mentioned

1. The Trade Desk
2. Scala (Programming Language)
3. Apache Spark
4. Andre Karpathy
5. Claude (Anthropic)
6. Spotify