How data leaders are hiring, securing resources, and thinking about AI

What should you be looking for when hiring data scientists? How do you make the case for more resources and headcount? What will AI mean for the data team?

These are the deliciously complex questions we explored at our event, The Data Blueprint, with three seasoned data leaders: Tuhin Ghosh, Head of Platform Science at Coinbase; Sumeet Marwaha, Head of Data at Brex; and Julia King, Head of Data at Gusto.

As data teams get more demands, and AI continues to evolve, it's important to continually evaluate how you measure, and communicate value in ways that matter to your organization. The answers, as our panelists revealed, are equal parts science and art. Here are the highlights from the conversation.

Hiring data scientists: curiosity and business framing are at the top of the list

"What I care most about is curiosity and passion for data," says Sumeet Marwaha, Head of Data at Brex. "If you don't have that, along with technical SQL skills and understanding of mathematical concepts, you're never going to get there."

This sentiment was echoed by Julia King from Gusto, who takes a similar stance: "I think the biggest value that the data person can bring to a team or project is curiosity, asking questions, and being open-minded." Her litmus test during interviews? "Are they hearing your request but thinking about what the underlying problem is that somebody is trying to solve?"

Tuhin Ghosh from Coinbase also emphasized the importance of business framing: "When I'm looking at an L6 staff data scientist, I'm looking at somebody who can frame problems. That wasn't the case three years back."

In sum, technical skills remain important, but true differentiation comes from candidates who can navigate ambiguity, reframe problems, and bring genuine curiosity to the table.

Justifying data team tools and headcount

How can you help justify the tools and headcount of your data team? "We've been trying to explain how in the real world, if we don't hire these people to build the foundation or invest in data quality – like, how long would it take for us to build that risk model?" Gusto’s Head of Data, Julia King says.

The case for data quality investments has become more compelling in the AI era. As Julia points out, "When we had humans building models, a lot of the complexity was hidden." But now, with AI tools, if the data underneath is not clean and well-structured, they can quickly break down. "As a leader, you start realizing that the value of that data foundation.”

Brex's Head of Data, Sumeet Marwaha, also uses proof-of-concepts to earn buy-in down the line for tools and headcount: "With a proof of concept... you want to show before you get that headcount that it's going to be worth it." He suggests stacking projects back-to-back to demonstrate sustained value rather than one-off needs: "It's not just 'Oh, I have a new project, I need someone for this.' It's 'what's going to happen next quarter?'”

The headcount conversion test: A data leader's litmus test for true need

"Are you willing to convert an existing headcount into a data position?" asks Sumeet, Head of Data at Brex. "That's a good test: How important is data to you, really?"

When resources are tight, this simple question cuts through aspirational requests to reveal genuine data needs. Rather than automatically adding data resources upon request, Sumeet challenges departments to demonstrate their commitment.

The finance team provided a perfect case study: "We were seeing lots of spreadsheets passed around with numbers that should be in a table somewhere," Sumeet explains. "At some point, you ask: should we hire another financial analyst or data engineer to automate these manual processes?"

The decision point becomes clear when inefficiencies pile up: "When you start seeing the exhaust of non-data person's work accumulating, that's when it becomes obvious." This observation led to a conversation with Brex's CFO and ultimately resulted in securing a new dedicated data resource.

AI won’t replace your team: it will help them focus on more valuable work

Coinbase’s Head of Platform Science, Tuhin Ghosh, references Jevons’ Paradox in his metaphor about AI. "Most people think broadening freeways reduces traffic, but it actually increases it," says Tuhin. "The same is true with AI and data teams."

Contrary to the myth that AI will reduce the need for data scientists, Tuhin sees it as an opportunity to tackle work that's been waiting in the wings. "With AI taking on the constant stream of ad hoc stakeholder requests that overwhelm most data teams, we can finally focus on higher value-add work," he explains. "In fintech, where regulatory landscapes shift rapidly, my vision is for us to dive deeper into experimentation, causal inference, and offline policy analysis — specialized work we simply haven't had the bandwidth for."

Focusing on people, process, and possibility

As data teams continue to juggle more demands, our panelists remind us about the importance of elevating the basics. When it comes to hiring, it’s not just about bringing on the “perfect” technical wizard — candidates should demonstrate curiosity and understand the business context behind every question. When making the case for more headcount, demonstrating a proof-of-concept can go a long way — as does citing the importance of investing in data quality with the rise of AI. And, when it comes to AI, the general consensus is that, over time, it will free up the data team to do more strategic work — making now the time to prepare your data.

Watch the full talk here.