Citi hires Google's director of engineering, shuffles AI leaders

• Key insight: Citi has hired Google's director of engineering to oversee development of agentic AI frameworks and applications.
  
• What's at stake: The hire and other AI leadership moves come at a time when many U.S. banks are trying to deepen their AI leadership benches and add management around agentic and generative AI projects.

Citi has hired Sebastian Gnagnarella, an AI leader from Google, and laterally moved other executives into its central AI team, according to an internal memo seen by American Banker. It's a sign of the times — U.S. banks have been hiring AI talent from tech companies as the need for expertise in agentic and generative AI management becomes more pressing. The reverse is also true: they've been losing AI leaders to tech companies like Google and Microsoft.
David Griffiths, who was previously Citi's CTO and was named group head of AI a few months ago, made the internal announcement.

Processing Content

Griffiths' team leads the firm's strategy, governance and approach to AI. The bank also has leaders who oversee AI within business lines and functions. Citi did not say whether the recent moves create a net increase in AI leaders.

In the memo, Griffiths noted that Citi has rolled out enterprise AI tools to more than 180,000 employees in 87 countries. It's also applied AI to know-your-customer compliance and to client solutions like Citi Sky, an AI-based digital assistant and virtual wealth advisor launched by Citi Wealth.

"Our AI journey is well underway, and we have reached an important inflection point," Griffiths said. "As we move into the next phase, strong leadership will be essential to ensure we continue delivering AI solutions that enhance our competitiveness while upholding the highest standards of safety, responsibility and governance."

Gnagnarella joined Citi as head of applied AI and agent factory and will oversee development of agentic AI frameworks and practical AI applications, Griffiths said. Gnagnarella previously was director of engineering at Google, where he led the deployment of the Gemini Enterprise Platform to Google Cloud clients.

In March, Citi hired Brian Saluzzo away from Google to be Citi's chief information officer. At Google, he was vice president of core developer engineering and product management.

Citi has also lost talent to Google, Microsoft and other tech companies in recent years, according to Vijay Swaminathan, CEO of workforce planning and talent-intelligence company Draup. The bank has lost five AI and data science leaders to Microsoft and three to Google in the recent past.

Other changes Griffiths announced at Citi were lateral moves. Aida Filipovic was moved from Citi's central controls team in April to become the bank's head of responsible AI. Her team is responsible for developing an AI governance model, AI ethical principles and an AI command center.

Niamh O'Sullivan, who used to lead Citi's technology strategy in the chief technology office, is joining Griffiths' groupwide AI organization as head of AI strategy, where she and her team will set an enterprise roadmap for AI development, deployment and adoption.

Sheetal Chanderkar is transitioning from her current role as head of business execution and chief of staff for the chief operating officer organization to become chief operating officer for the Groupwide AI team.

The bank is hiring a head of AI-first development, who will help drive the buildout of AI-native products and platforms at Citi.

During Citi's third-quarter earnings call last week, CEO Jane Fraser said the bank is "approaching AI through four lenses: how it can help us grow; how we can deploy it to operate more efficiently; how we can teach our people to use these new capabilities; and how we can strengthen our defenses given the enhanced threat environment we and our clients face."

She said AI is accelerating the bank's product-development life cycle, so it's launching new products at lower cost and at faster speed.

"We're typically seeing about a 30% to 40% increase in the speed with which we are launching products," she said. About 14,000 engineers use Cognition's agentic coding tool, Devin.

"That makes it much more efficient for us to do the work in-house," Fraser said. "So we've been in-sourcing work. We've reduced our tech vendor spend by 30% year-over-year as a consequence, and that's still got a long way to run." Response times have improved for clients by 21%, she said, and client satisfaction scores are better.

Across the banking industry, AI hiring is up, according to an analysis conducted by Draup. AI-related job postings increased 49% year-over-year, from about 93,686 to 139,819, according to the company's recent analysis.

Draup's primary data sets include 800 million publicly available professional profiles and 700 million job postings. Its analysts looked at job openings of banking companies over the last 12 months and the prior 12 months to see how things have changed.

Read more:

• As fintechs nab bank charters, what's happening to BaaS?
• Overheated data center knocked 23 credit unions offline
• Chris Britt still says banking is broken, but Chime can fix it
• Credit unions bulk up a business that banks long dominated

Overall hiring in the banking industry has been flat or slightly declining, Swaminathan said.

"But what they're looking for is what's changed," Swaminathan told American Banker. "The ask, for AI-related skills, AI-related expertise, whether it is building or orchestrating or governing, has increased drastically while overall hiring has remained flat."

In general, there's been a decline in entry-level hiring in banking. "They're looking for experienced folks," Swaminathan said. "This means you have to find talent from your competitors — other banks, tech firms. So the squeeze is coming from a need for experienced talent that can build agentic systems, that can build AI models, that can orchestrate AI models, and that can also govern these models."

Last year, companies were hiring a lot for AI builders, he said. "They were hiring for data scientists, they were hiring for AI engineers and so on. But what we've also started realizing is building is insufficient because deploying is the bottleneck. Integrating things is the bottleneck. So deployment has become a big focus. Orchestrating all of this across different workforces has become a big focus, and that's where everything is shifting."