Senior AI PM for Data and Governance

About this role

SENIOR AI PRODUCT MANAGER, DATA, ANALYTICS, EVALUATION & GOVERNANCE Transform healthcare with us. At Qualified Health, we're redefining what's possible with Generative AI in healthcare. Our infrastructure provides the guardrails for safe AI governance, healthcare-specific agent creation, and real-time algorithm monitoring, working alongside leading health systems to drive real change. This is more than just a job. It's an opportunity to build the future of AI in healthcare, solve complex challenges, and make a lasting impact on patient care. If you're ambitious, innovative, and ready to move fast, we'd love to have you on board. JOB SUMMARY Qualified Health is seeking a Senior AI Product Manager to own the connected core of our platform: data, analytics, evaluation, and governance. That means the data layer as a product, the analytics and insights surfaces our customers see, the evaluation frameworks that tell us whether AI outputs meet the bar, the governance spine that makes all of it safe to run in healthcare, and the contracts that connect the data organization to every product we ship. These four are one system: data feeds the products, analytics measures them, evaluation proves they work, and governance makes them trustworthy at scale. This is an integrator and enabler role, not a control role. The data platform and analytics teams own their domains, their delivery, and their technical decisions; that does not change. What is missing today is the connective tissue: product-shaped data work is spread across a data platform team, an analytics and new-product team, and a platform engineering organization, and nobody owns the seams between them. You are that person. You make these teams faster by absorbing the coordination work that currently lands on their leads: writing the acceptance criteria before build, defining the contracts between teams, running intake and prioritization for analytics asks, and making sure what gets built once is reusable everywhere. When a care gap product needs a data pipeline, you make sure it is scoped once, built to generalize, and reusable for the next customer. When a dashboard metric ships, you make sure it is defined once, governed, and consistent everywhere it appears. When engineering and data disagree about who owns a layer, you are the person who has already written it down. Your success is measured by whether the data platform and analytics teams say you make them faster. If they route around you, the role has failed. You will sit on the Platform pod, reporting to the SVP of Product, and partner daily with the data platform lead, the analytics and new product development lead, engineering leadership, and peer product managers. WHAT YOU WILL OWN You own products, contracts, and processes. The teams own their domains. - The data layer as a product: The serving contracts, gold-layer marts, and semantic views that assistants, chat, workflows, and dashboards consume. Defined once, versioned, and stable enough that the data platform can refactor underneath without breaking products. The data platform team builds and owns the platform; you own the product definition of what it serves and to whom. - Analytics and insights products: Customer-facing dashboards, usage and adoption analytics, cost and token observability, and the KPI catalog. One definition per metric, a lightweight vetting process for anything customer-facing, and no metric proliferation. The analytics team owns the builds; you own intake, prioritization, and the catalog so requests stop arriving from every direction at once. - Data product pipelines for clinical products: The data and scoring pipelines behind products like Care Gap Optimizers: acceptance criteria written before build, generalization requirements set at the start, and clear seams between data, AI engineering, and application engineering. - Evaluation as a product: The frameworks, datasets, and gates that determine whether an AI output is good enough to ship and stays good enough in production: validation criteria before build, generalization requirements at full population scale, and post-go-live monitoring. Evaluation stops being a per-team improvisation and becomes shared platform capability. - Governance as a product: The controls and evidence that make AI safe to run in healthcare: what is monitored, what is auditable, what a customer's compliance team can be shown. Governance is the platform's spine and its differentiator; you make it a product surface, not a checklist. - Team Integrations and Hand-offs: Written contracts for who owns ETL, business logic, evaluation, and serving across the data organization and platform engineering, so process questions are settled in documents instead of escalations. KEY RESPONSIBILITIES - Own the roadmap for data and analytics products, balancing customer commitments, internal builder needs, and platform reuse - Map every analytics question to its source system (product databases, event analytics, observability tooling, the lakehouse) and own the architecture decisions for how data lands in the serving layer: tables, granularity, fields, cadence - Turn per-customer data builds into reusable data products with clear interfaces, so the second customer costs a fraction of the first - Define and enforce generalization requirements for scored and modeled outputs: what was validated on hundreds of patients must hold at hundreds of thousands - Write complete acceptance criteria before work commits to a sprint; no ambiguity reaches engineering or data - Run the vetting process for customer-facing KPIs and own the metrics catalog end to end - Own cost observability as a product: every model call attributed, every workflow priced, actuals replacing estimates - Define the evaluation gates for AI-powered products (validation, generalization at scale, post-go-live monitoring) and own them as reusable platform capability rather than per-pr

View the original posting ↗