A Goldman Sachs partner in technology shares the skills young job seekers need in the AI workplace

• Bracha Cohen, a partner and engineer, said young talent should focus on how systems work.
• She advised computer science majors to learn how to ask good questions — of humans and machines.
• Goldman Sachs hired a record low percentage of interns last year.

Bracha Cohen has a front row seat to Wall Street's AI revolution — and to how young people can compete in it.

"I would tell the new generation of graduates, in this world where AI is so transformational, to build judgment and not just skills," Cohen, a partner within asset and wealth management engineering, said. "AI may automate execution, but it can't fully replace decision-making, systems thinking, and ethical reasoning."

Cohen joined Goldman as a programmer in 1994, long before anyone had to prove AI fluency on their applications. She said that serving in various roles across business lines helped her ascend to partner, the firm's top leaders.

Today, her engineering team in asset management focuses on automating operations to help the business scale, including through AI. As of now, the booming business — which holds a record $3.6 trillion in assets — uses AI for routine work, like analyzing and summarizing data, Cohen said.

As white-collar hiring slows and anxiety about AI in junior roles grows, Cohen said young engineers should focus less on simply completing tasks and more on how systems function. Mastering engineering fundamentals is key, she said, since AI should serve "as leverage, but not as a crutch."

She added that computer science majors should practice evaluating risk and crafting good questions, both for other people and AI models. Two other Goldman partners also previously said that interpersonal skills and communication are becoming increasingly crucial in the AI workplace.

And engineers who want to work on AI in particular have their own set of criteria. Dan Popescu, a newly promoted managing director and Goldman's head of AI engineering for asset management, previously told Business Insider that the most competitive hires need a suite of skills: knowledge in AI engineering, finance, and traditional software engineering.

Goldman spent $6 billion on technology last year and has rolled out internal AI tools, including an assistant and a limited banker copilot. In an October memo, the firm laid out the latest phase of its OneGS initiative, which it says will drive efficiency, slow hiring, and create a "limited reduction" in roles.

CEO David Solomon is one of several big bank leaders who have said that, in the long run, AI won't reduce head count, and that the firm needs to focus on attracting more high-quality talent.