Principal Product Manager (AI)

Craniometrix Using AI to help dementia patients age at home Principal Product Manager (AI) $180K - $200K, equity 0.10% - 0.50%, New York, United States Job type : Full-time Role : Product Experience : 6+ years Visa : US citizen/visa only About the role Craniometrix delivers Care Navigation for patients and their families needing support with complex long-term conditions - starting with dementia. Having launched our services in summer 2025 we have already raised a $15m Series A and are providing care for 3,000 families. We’re on track to support over 17,000 families by the end of 2026. Our care navigators are the heart of our service, connecting regularly with patients and families. Your job is to build the AI frameworks and internal tools that eliminate administrative friction, maximize clinical accuracy, and deepen personalization. By ensuring our AI has the high-fidelity context it needs—and the safety guardrails it requires—you will empower our team to focus 100% of their energy on supporting families. This is a senior IC role for a builder who sits at the intersection of Operations, Product, and Engineering. You will report to the VP of Operations and be responsible for the internal AI frameworks that power our staff-facing tools. You aren't just a PM writing specs—you are a systems architect. You will define the logic and safety rules for our internal AI, and then use our AI codex and low-code infrastructure to help build the tools our care navigators and operations managers use every day. What You’ll Do Build the Navigator Workbench: Design and ship internal AI tools that help our care navigators summarize patient history, deliver more personalized care, flag risks, and automate Medicare-compliant documentation. Optimize the Data Feed: You will be a key stakeholder in our data collection process. You’ll provide a critical feedback loop to the team on how we capture data—improving transcript quality, refining form inputs, and optimizing assessment tools to ensure the AI has high-fidelity context. Define the Rules & Frameworks: Develop the 'logic engine' for our internal AI. You’ll create the prompts, decision trees, and guardrails that ensure our staff receives safe, reliable, and unbiased support. Bridge to Engineering: While you will build extensively using our AI codex, you will work closely with the core Engineering team to ensure your internal tools integrate deeply with our data architecture and production stack. Architect the Operational Sandbox: Build the underlying frameworks—including data schemas and logic evaluators—that allow non-technical operational leads to contribute to the build process, effectively turning your architectural work into a force multiplier for the entire department. Operational Safety & QA: Own the monitoring of AI outputs. You will build the frameworks to audit AI suggestions, ensuring every piece of support provided to our navigators meets our high quality bar. Efficiency Architecture: Constantly identify "friction points" in the navigator workflow and build AI-driven solutions to remove them, while managing model costs and latency. Who You Are The 'Systems' PM: You have 5–7+ years of experience in healthcare or high-complexity operations. You think in workflows, edge cases, and feedback loops. Data-Obsessed: You understand that 'Garbage In = Garbage Out.' You are proactive about identifying when a lack of context or poor data quality is hampering AI performance and you know how to fix it at the source. Hands‑on AI Builder: You have 2+ years of experience building with LLMs. You are skilled in using AI to build and don't need a dev team to move an idea from 'concept' to 'internal tool.' Internal-Tool Obsessed: You find joy in building tools that make it easier to deliver care. You care about utility and safety above all else. Pragmatic Engineer‑Lite: You understand enough about software architecture to partner effectively with engineers, but you prefer the high‑velocity environment of building in the Ops layer. About Craniometrix Craniometrix has raised over $21 million to help make dementia suck less. Dementia is tough – but it can be easier. Turns out, if you give families the right non-clinical coaching and support, you can avoid 30% of hospitalizations for these patients. And Medicare has recently started paying for that coaching. We launched our care navigation service for dementia patients in July of 2025, hit $4M of ARR, and just raised a $15M Series A. We leverage AI to provide best‑in‑class care, alleviating the burden on providers. Come help us build! #J-18808-Ljbffr