Product manager in AI

Spark, our AI agent for product teams, exists to change that: to turn that volume into the handful of insights behind the decisions a product leader has to make and defend. We're looking for a Principal Product Manager to own that opportunity end to end: a hands-on senior IC from an AI-first product company who owns the pipeline that turns raw signals into reliable insights, the agent behavior that reasons over them, and the quality of the answers and artifacts Spark produces, and who partners directly with our founders on the hardest problems along the way.

• Know the customer better than anyone. Understand the questions product leaders most need answered, how modern product organizations actually work, and where an agent removes work rather than adding tooling. Bring that to every prioritization call.
• Own the AI Insights roadmap. Set the strategy, roadmap, and metrics for the domain. Decide which questions Spark answers first, what a great answer looks like for each, and the quality bar it meets before it ships. Lead the hardest problems in the domain, like telling a feature gap from a use case we've never served, or discovering segments the CRM has no field for.
• Own the path from signal to decision. Decide what data and context Spark needs to answer reliably, how it flows through the data model, retrieval, agent harness, and experience, and whether a weak answer gets fixed through better data, retrieval, prompting, tooling, model behavior, or UX. Decide what belongs in Insights and what belongs in the shared AI platform.
• Define how agents behave and how answers show up. Decide what Spark does in each context, how it handles uncertainty, and when it asks for help. Decide when an answer is conversational, structured, visual, or fully interactive, and lead the move from text outputs to rich, interactive artifacts.
• Own quality and evals. Write the evals that define a good answer, insight, or artifact, and codify them so engineering has something precise to build against. Read the traces yourself, watch for regressions, and make sure every release makes Spark measurably better.
• Anticipate the frontier. Track what the next generation of models and agent frameworks makes possible, position Spark to capture it first, and give GTM teams a crisp articulation of what the agents can and cannot do.
• Prototype in code by default. When words aren't settling a debate, build the agent with AI coding tools and put it in front of customers.
• Raise the bar. Be the first user of Spark for your own work, and mentor other PMs on AI product craft through pairing and example.

• 8+ years in product management, all of it hands-on. A senior individual contributor on complex products, with meaningful time in enterprise B2B SaaS, direct ownership of an AI or LLM-powered product in production, and an AI product that is visibly better because you owned it.
• Deep expertise in designing and developing agentic AI products. You've shipped products where LLM-based agents plan, use tools, and take multi-step actions on the user's behalf, you understand intimately how and why they fail, and you have real opinions on where agents belong in structured enterprise software and which parts of product management are worth rebuilding around them.
• Systems thinker who has built AI on real customer data. You've owned the data and retrieval pipeline behind an AI product, not just the prompt. You reason comfortably about data models, pipelines, and architectural trade-offs, you understand how LLMs behave in practice, including context, tool use, latency, cost, and non-determinism, and you read code, model outputs, and traces with confidence.
• Evals are part of your craft. You've built and operated evaluation frameworks, offline evals, online metrics, and human review loops, and used them to drive product decisions rather than just report on them. You can show us prompts and evals that shipped.
• Strong design taste. You know when an answer should be a sentence, a table, a chart, or an interactive artifact, and you hold a high bar for the final experience, not just the model behavior.
• AI-first in how you work. AI agents are how you research, synthesize, analyze, and prototype every day. When a debate needs settling, you build it with Claude Code, Cursor, or whatever fits, and you use the newest models and agents the week they ship.
• Founder mentality, low ego. You own outcomes end to end, turn ambiguity into clear decisions in a space where best practices are still being written, stand up for what's right with challenging stakeholders, and work in the open with engineers and designers as peers. You can turn a subtle finding about agent behavior into a clear decision for an executive, an engineer, and a GTM team, and each leaves clearer than they arrived.
• Candidates must have the legal right to work in the EU.

• Shape the full stack of a category-defining AI product for product teams, not a thin layer on top of a model: from the data model and feedback pipeline through the agents to the experience, in a category being defined right now.
• Work directly with our founders, CEO Hubert Palan and Chief AI Officer Daniel Hejl, on the company's biggest AI bet.
• Real scope, real ambiguity, real accountability, with the mandate and executive partnership to actually move the product.
• Operate the way the industry's best AI-first product teams do, with the tooling, velocity, and autonomy to match.
• Build for product makers, the most demanding, product-literate audience there is. When it works on them, it works.
• 5 weeks of paid time off per year
• 9 paid sick days per year, also for mental-health days
• 1 paid Volunteer Day per year
• Stock options
• USD 300 per year professional development stipend
• Mental wellness program with online therapy sessions
• Carrot: fertility, adoption and family-forming support for you and your partner (pregnancy, postpartum and menopause support in the UK and Ireland)
• One-off home-office stipend (amount by location)
• Monthly fitness or wellness support (form by location)
• Free 1Password Families account
• Annual company retreat and team off-sites
• Private medical insurance with Vitality
• Scottish Widows workplace pension: 5% employer and 4% employee contribution, auto-enrolment
• Life insurance (death in service) with Canada Life
• Income protection with Canada Life: 66.67% of salary, payable after 26 weeks
• 2 weeks of paternity leave at full pay
• Gym stipend: GBP 42 monthly gross allowance
• One-off home-office stipend of GBP 400
• Multisport card, with options for a +1 and children
• Catered lunches in the office every working day, always with a vegetarian option