We Didn't Invent the Product Engineer. We Just Made Time For One.

Reflections | Business Evolution

References: AI Changed Engineering. Hiring Hasn’t Caught Up.

Everyone's talking about "product engineers" like it's a 2026 invention. It isn't.

Sherif Mansour, a Distinguished Engineer at Atlassian, coined the term back in 2018. His definition was simple: an engineer who thinks like a PM — someone who pushes on the why, not just the how. Jean-Michel Lemieux picked up the label and ran with it internally. Nothing about that idea needed AI to exist.

So why does it feel like a brand-new category now?

Because the constraint that used to make the product-engineer hybrid rare — time — collapsed. The role exists because AI has collapsed the time between idea and working prototype from weeks to hours, and that collapse rewards generalists who can hold product judgement, engineering, UX taste, and data instincts at once. When one person can prototype what used to take a team a sprint, the bottleneck stops being execution and starts being judgement. (Turing College)

That's the real shift. Not a new job title — a redistribution of where the bottleneck sits.

The Pragmatic Engineer's 2026 engineer survey offers a cleaner lens than most of the hype pieces. It splits engineers into three archetypes: Builders, who care about craft, architecture, and quality; Shippers, who focus on product outcomes, features, and experimentation — a category that includes product engineers and previously hands-off managers; and Coasters, who get work done without much taste or concern for quality. (The Pragmatic Engineer)

What's more interesting than the labels: engineer and engineering-manager roles are becoming more similar — engineers now orchestrate and context-switch more, while managers get more hands-on. The org chart is flattening from both directions at once. (The Pragmatic Engineer)

I'd be more skeptical of this if it were just a wave of think pieces. It isn't. 2026 AI hiring data groups roles into distinct categories with separate talent pools and comp curves — and "applied AI / LLM product engineers," hired specifically to ship AI features customers actually use, is one of the largest. Ramp is hiring for it at every level. Steampunk's posting reads almost like a job description someone wrote after reading Mansour's original post. (Recruits Lab)

No — and be wary of anyone telling you it is.

The "kill your specialists" framing is opinion dressed up as inevitability. The more honest version of the argument is a genuine trade-off, not a verdict: coordination cost versus depth. Teams that collapse handoffs move faster until the thing they're building needs the kind of depth a generalist doesn't have time to develop.

A few things I'd bet on instead:

• Specialists don't disappear — they get scarcer and more expensive at the edges (security, infra, ML systems)

• Most early-stage teams will keep optimising for shipping speed over architectural depth, because that's the trade they can afford to make

• The convergence Pragmatic Engineer points at — engineers and managers looking more alike — will outlast the "product engineer" label itself

The title will change again in two years. The compression won't reverse.