Senior Product Manager, AI

About the role

This role owns the layer of AI quality, measurement, and platform strategy that sits behind every AI experience delivered to members and internal teams at WHOOP. You will define how AI output quality is evaluated across the business, shaping the datasets, methodologies, and scoring frameworks that make trust continuous rather than anecdotal. You will partner closely with AI engineers and research scientists to translate model behavior into reliable member experiences across every surface. You will product manage the internal AI development platform, lowering the barrier for teams to ship AI features safely and at scale. You will balance emerging research techniques against unit economics, deciding which capabilities to pursue as scalable bets versus experimental side projects. You will act as the central evaluation and cost discipline layer, ensuring AI initiatives pay for themselves in quality, efficiency, and member impact. This is a platform-oriented role that sits at the intersection of data, engineering, and product, requiring comfort with ambiguity and rigor in execution.

Key facts

What you'll do

• Shape how AI output quality is measured across WHOOP's AI experiences, including the evaluation methodology, datasets, and scoring that keep quality continuous and trusted.
• Build the evaluation habit across the company by making evals self-serve and ensuring teams shipping AI features actually run them before launch.
• Help decide which models power which experiences, weighing quality, latency, and cost as the model landscape shifts and new providers emerge.
• Product-manage our internal AI development platform and drive its adoption so that anyone at WHOOP building with AI has a fast, well-instrumented path from idea to shipped.
• Track the unit economics of AI at WHOOP, including cost per interaction, and drive the tradeoffs between quality, latency, and spend.
• Explore where AI capability goes next, from fine-tuning and reinforcement learning to distillation and emerging foundation models, turning a broad field of options into a small number of well-scoped bets.
• Partner with research and engineering as a peer, translating model behavior and constraints into clear requirements and member-facing outcomes.
• Define and own the guardrails and tooling that enforce evaluation standards, enabling teams to move quickly without compromising quality.
• Prioritize evaluation workstreams by impact on member experience and business risk, balancing experimentation with reliability.
• Work cross-functionally with analytics, research, and operations to align AI initiatives to company goals and regulatory considerations.
• Maintain a living view of the AI roadmap, surfacing dependencies, risks, and progress to stakeholders at all levels of the organization.
• Champion best practices in prompt engineering, data hygiene, and model observability to raise the baseline of AI quality across WHOOP.
• Identify and close gaps in evaluation coverage, ensuring that new features are backed by meaningful metrics before release.
• Represent the AI product perspective in hiring, roadmap, and vendor decisions, influencing long-term strategy for internal tools and external providers.
• Translate complex model updates into clear product decisions, helping stakeholders understand tradeoffs and expected outcomes.
• Maintain a lightweight but rigorous decision framework for whether to build, buy, or partner with AI capabilities.

Requirements

• Experience shipping AI or machine learning products where the underlying capability was probabilistic rather than deterministic.
• Depth in evaluation and measurement, having defined how something gets measured, not only reported on it.
• Working literacy in modern ML, including what fine-tuning is, the main flavors of it, what RL-based post-training is for, and how the current generation of techniques fits together.
• Experience with platform or internal-facing products where the users are other teams and success depends on enabling those teams.
• Technical fluency sufficient to partner with research and engineering as a peer and translate model behavior into member experience.
• A track record of driving cross-team programs with many dependencies and no direct authority.
• Sound business judgment, including the ability to quantify tradeoffs, reason about cost, and decline work that will not pay for itself.
• Comfort operating in ambiguity, with evidence of turning a broad set of possibilities into one well-scoped bet.
• Strong commitment to embracing and leveraging AI tools in day-to-day tasks, ensuring AI-assisted work aligns with the same high-quality standards as personal contributions.
• Willingness to relocate to Boston, MA, as required for in-office work at the WHOOP office.

Nice to have

• Internal AI tooling or evaluation frameworks.
• Experience partnering with a research organization.
• Wearables or sensor data.
• SQL or light scripting.

Practical notes

This role is based in the WHOOP office located in Boston, MA. The successful candidate must be prepared to relocate if necessary to work out of the Boston, MA office.

Interested in the role, but don't meet every qualification? We encourage you to still apply. At WHOOP, we believe there is much more to a candidate than what is written on paper, and we value character as much as experience. As we continue to build a diverse and inclusive environment, we encourage anyone who is interested in this role to apply.

WHOOP is an Equal Opportunity Employer and participates in E-verify https://www.e-verify.gov/to determine employment eligibility. The WHOOP compensation philosophy is designed to attract, motivate, and retain exceptional talent by offering competitive base salaries, meaningful equity, and consistent pay practices that reflect our mission and core values. At WHOOP, we view total compensation as the combination of base salary, equity, and benefits, with equity serving as a key differentiator that aligns our employees with the long-term success of the company and allows every employee to share in our growth.