Staff Product Engineer

Superads is building AI agents for marketers — systems that don't just surface data but actually think through creative strategy, analyse performance, and help teams make better decisions before a single asset goes into production. We're moving fast in a space that's genuinely wide open, and we're looking for a Staff Product Engineer to help us build the product layer on top of that ambition.

We're not looking for a consultant who's mastered a particular layer. We're looking for an engineer who's restless until the product is right.

• * Build the interfaces through which marketers interact with AI agents — conversational flows, agentic workflows, and UI that makes complex reasoning feel simple and trustworthy.
• Own the product layer end-to-end: from the activation flows that bring new users in, to the interaction patterns that make them come back, to the feedback loops that help the AI get better over time.
• Figure out the right patterns for making an AI strategy layer legible and useful inside a modern SaaS product — there aren't established playbooks here, and you'll be one of the people writing them.
• Go full-stack when the problem requires it — into the API, the data layer, or the AI integration — without creating drag for the rest of the team.
• Bring a PM-level perspective to growth: identifying which flows are losing people and why, forming opinions on what healthy usage looks like, and changing what you're building when the data says you should.
• Represent product and design interests across teams, pushing back on decisions when the bar is slipping and being right often enough that people listen.

• * 8+ years in high-performance product teams at scaleups — somewhere the engineering culture was set by engineers, not handed down as a process.A strong home base in frontend (Next.js [http://next.js/] / React) with demonstrated ability to go full-stack when the problem demands it — you follow the work, not your lane.
• Real experience building AI-powered product surfaces in production: agentic systems, LLM integrations, conversational interfaces, or retrieval-augmented workflows — and the product instincts to make them feel trustworthy rather than just functional.
• Experience evaluating and improving AI output quality — whether through evaluation frameworks, feedback loops, or iterating on prompts and retrieval. You've thought about what good looks like when the system is reasoning, not just retrieving.
• You think about activation and retention the way a PM would. You've looked at usage data and changed what you were building because of it — not as a process, but because it bothered you that people weren't coming back.
• Product and design intuition strong enough to represent those interests when the PM or designer isn't in the room. You've pushed back on bad product decisions more than once and been right often enough that people listen.
• You don't need to be managed. You take ownership of outcomes, not tasks, and you're honest about what's working and what isn't.