The opportunity
Klaviyo's AI & Analytics pillar builds the intelligence layer that helps marketers understand their customers, personalize every interaction, and prove the impact of their work. Two of our highest-leverage products live at the center of this: Product Recommendations and Audience Optimization.
What you'll do
Own the roadmap for Product Recommendations, including model type selection,: catalog data structure decisions, and enterprise-grade customization such as filters, guardrails, and exclusions
Own the roadmap for Message Prioritization, defining how the model balances: message frequency limits against conversion and revenue outcomes
Define what model success looks like for each feature area and build the: frameworks to measure lift, conversion improvement, and revenue impact over time
Partner closely with ML modeling teams and software engineering teams to: translate model capability into features customers actually trust and adopt
Design product experiences that make ML-driven decisions understandable and: explainable to marketers, so the model never feels like a black box
Influence teams beyond your direct authority to align on priorities and: deliver feature outcomes, consistent with how Klaviyo's product orgs operate
What they're looking for
- –5+ years of PM experience owning ML-powered or data-driven product features: in a production environment, not just AI-adjacent products
- Direct experience evaluating and tracking ML model performance: you understand lift measurement, experiment design, and what good model quality tracking looks like
- Comfortable partnering with both software engineering and ML modeling or data: science teams. You speak enough of both languages to make good calls without needing everything explained to you
- Genuine customer obsession: you understand why a brand decides to pay for an ML feature, and you translate that understanding directly into roadmap decisions
- Strong instincts for human behavior and UX: you know how to design trust, clarity, and adoption into a feature that could otherwise feel opaque or unpredictable
- Comfortable working with complex data structures, like product catalogs, and: turning that complexity into simple, understandable customer-facing controls
- High agency: you take ownership of outcomes, not just deliverables, and you: pull in the partners you need even when they don't report to you
- A track record of driving cross-functional alignment and shipping through: ambiguity, without relying on direct authority to get it done