The staff engineer role has not been diminished by AI, it has been redefined in a way that demands more, not less, from its senior practitioners. Jordan Cutler, Senior Staff Engineer at Pinterest, made this clear at QCon San Francisco 2026 when he described how the job changes once every engineer on the team has access to the same AI tools. His presentation cuts through the noise around AI replacing roles and instead frames a practical challenge: when everyone can generate code, what distinguishes the engineer who shapes the system?
This is the question that should matter to anyone responsible for architecture, mentoring, or technical strategy. The answer, based on Cutler's observations, is that staff engineers must now focus on problems that AI cannot solve alone. A model can produce a function, but it cannot negotiate cross-team dependencies, evaluate long-term tradeoffs, or build the shared context that makes a codebase coherent. As teams adopt tools like those explored in Connect Claude to your trading account for smarter portfolio reviews, the staff engineer's value shifts from writing more code to deciding which code should exist at all. The same principle applies: the tool is only as useful as the judgment guiding its use.
We see a direct parallel in how Vercel's new SDK is reshaping frontend development. When Explore how Vercel's new UI SDK helps agents build accessible interfaces, the engineer's role becomes less about implementing every interaction and more about defining the interaction model itself. Cutler's talk suggests that staff engineers at Pinterest are experiencing the same shift: their primary output is no longer code commits but architectural decisions, review processes, and the standards that keep a team aligned. The concrete takeaway is that promotion criteria for senior technical roles should be rewritten now. If your organization still evaluates staff engineers primarily on individual output, you are measuring the wrong thing.
What remains unresolved is how teams will develop the judgment that AI cannot replicate. Cutler's presentation implies that staff engineers must become deliberate teachers, embedding their reasoning into documentation, design reviews, and pair sessions so that the rest of the team, and the AI tools they use, operate from a shared understanding. That is a specific, difficult skill that most engineering career ladders do not yet reward. The open question for leaders is whether their performance frameworks can adapt fast enough to catch up.
