Job Overview
About Abridge Abridge was founded in 2018 with the mission of powering deeper understanding in healthcare. Our AI-powered platform was purpose-built for medical conversations, improving clinical documentation efficiencies while enabling clinicians to focus on what matters most—their patients.
Our enterprise-grade technology transforms patient-clinician conversations into structured clinical notes in real-time, with deep EMR integrations. Powered by Linked Evidence and our purpose-built, auditable AI, we are the only company that maps AI-generated summaries to ground truth, helping providers quickly trust and verify the output. As pioneers in generative AI for healthcare, we are setting the industry standards for the responsible deployment of AI across health systems.
We are a growing team of practicing MDs, AI scientists, PhDs, creatives, technologists, and engineers working together to empower people and make care make more sense. We have offices located in the Mission District in San Francisco, the SoHo neighborhood of New York, and East Liberty in Pittsburgh.
The Role
Abridge's has multiple products - the core product is ambient clinical notes, but we’ve expanded surfaces: billing, clinical decision support, orders, nursing, etc. Measuring quality reliably and iterating fast without breaking clinician trust sit at the core of every single product’s success.
The Evals team owns the tooling, templates, and consultation that product teams use across the eval lifecycle: curate, build, iterate, and deploy. This builds the strategy, platform, and process that enables evals to be fast, trustworthy, consistent, and a true moat.
We are hiring a Product Manager to lead this platform. You will work with 10+ teams and partner closely with engineering, ML, data science, and clinician science. You will help set the standard for how Abridge measures AI quality, decide what the platform builds, and make evaluating a new model cheap and routine as frontier models ship frequently.
What You'll Do
- Drive product strategy and execution for the evals platform. Own the roadmap across the eval lifecycle and own outcomes against it.
- Build the shared measurement infrastructure. Help build the systems that let any pod define quality, run experiments, compare models, and watch production. Own the standards for LLM judges, rule-based evaluators, human annotation, and online monitoring, and be clear about where each belongs.
- Make model selection fast and routine. Give teams a repeatable way to evaluate a new frontier model within days of release, and a defensible framework for when a post-trained specialty model is worth it over a prompted frontier one. Keep the eval system model-agnostic so it stays a neutral referee.
- Own the operating model across pods. Define the eval gates from early build through GA and steady-state monitoring, which gates are hard versus advisory, and who owns non-negotiable floors like critical-error rates. Land this acr