Banking giants turn to artificial intelligence as junior roles face existential threat
In the sunless cubicles of lower Manhattan, where bankers once worked 80-hour weeks to polish legacy code and prepare market research for Monday meetings, something quietly, perhaps irreversibly, shifted this summer. That something is Devin — a tireless AI software engineer that doesn’t sleep, doesn’t complain, and never forgets a line of code.
Goldman Sachs, a venerated titan of global finance, has embraced a London-born AI developer named Devin, created by Cognition Labs, to begin reshaping the digital soul of its vast operations. The rollout is not a small experiment. According to Chief Information Officer Marco Argenti, Goldman has already deployed hundreds of these AI coders — and plans to scale that number into the thousands.
This move, bold even by Wall Street standards, is less about replacing humans outright and more about redefining who (or what) does the work. Still, whispers across the City and Lower Manhattan echo the same concern: Is this the end of entry-level finance?
A Machine That Never Asks for a Bonus
Goldman Sachs employs over 12,000 human developers. That number once symbolized the firm’s commitment to in-house innovation and data-driven edge. But in Devin, the bank has found an upgrade that speaks fluent Python, commits bug-free updates, and completes tasks overnight — all without performance reviews.
“Devin will function like our employee,” Argenti told CNBC. “He’s going to start doing stuff on behalf of our developers.”
The “stuff” he refers to, for now, includes mundane but essential tasks: updating legacy codebases, running back-end scripts, and managing compliance tweaks across sprawling fintech stacks. Human developers, it’s implied, will then be freed up to think bigger, faster — or simply think less about the code and more about the strategy.
For Goldman, this isn’t just automation. It’s optimization.
Why This Isn’t Just Another AI Tool
AI has long flirted with finance — from algorithmic trading to robo-advisors. But Devin is different. Rather than replacing a function, he replaces a role. That nuance matters deeply, especially to junior software engineers and analysts hoping to climb Wall Street’s fabled corporate ladder.
Devin can read, write, and execute complex coding tasks with a precision that surpasses traditional developer assistants. He doesn’t hallucinate, his creators claim. And unlike other tools, he’s being positioned as a true colleague.
This positions Devin less like a hammer and more like a new pair of hands on the team. Only these hands don’t get carpal tunnel.
“We’re not replacing developers,” one Goldman VP said under condition of anonymity. “But let’s be real — we’re reducing how many we’ll need in the future.”
Human Labor on the Line
Goldman Sachs’ embrace of Devin comes amid a broader reckoning in the financial sector. Bloomberg Intelligence estimates that as many as 200,000 jobs could disappear globally from banking in the next 3–5 years due to AI.
Entry-level roles are most at risk. These include positions historically seen as proving grounds — analysts who clean Excel sheets, junior developers maintaining outdated systems, research assistants drafting investor memos.
| AI Impact Forecast in Banking (2025–2030) | Estimate |
|---|---|
| Global banking jobs at risk | 200,000 |
| Entry-level roles affected | 70% |
| Tasks likely to be automated | Routine coding, compliance checks, data entry |
| Average productivity boost from AI coders | 3x to 4x |
For aspiring graduates from MIT to Mumbai dreaming of a seat on Wall Street, Devin’s arrival isn’t a curiosity — it’s a warning.
The Long Game: Trust, Speed, and Control
Internally, Goldman’s leaders are keen to frame this AI transition as a partnership rather than a purge. The bank insists that AI will empower human workers, not erase them. But industry veterans remember similar assurances during the 2008 outsourcing boom — and many of those jobs never returned.
Still, Goldman’s ambitions are clearly futuristic. The firm isn’t looking at AI as a plugin, but as a foundation. Executives say they’re building systems that combine the predictability of AI with the creativity of experienced developers. It’s a hybrid model, but one increasingly tilted toward automation.
Three core motivations drive this push:
-
Productivity Leap: One Devin can do the work of multiple junior coders overnight, with consistent results.
-
Cost Efficiency: No salaries, no vacations, no HR overhead.
-
Speed to Market: AI tools deliver updates, patches, and innovations faster than any human team.
Not Just Goldman: A Trend Becomes a Standard
Goldman is not alone in this shift. Rival banks like JPMorgan Chase and Citigroup are already experimenting with their own AI development stacks. JPMorgan has quietly ramped up hires in its Applied AI division, while Citi has partnered with Palantir to embed predictive modeling in compliance systems.
What’s different with Goldman is the scale and transparency of its move. It’s not positioning AI as a backend enhancer. It’s making AI a face of its engineering workforce.
A Different Kind of Internship
For young coders graduating into this reality, the path forward is murky. The role of junior developer — once a rite of passage at investment banks — may become an endangered species. Hiring managers increasingly look for “AI supervisors” rather than pure programmers.
Argenti says tomorrow’s Goldman developer must understand how to work with Devin. The job is no longer about typing out code. It’s about steering intelligent agents who can write it better.
In Wall Street’s machine-driven tomorrow, Devin isn’t coming. Devin is already here.
