Staff ML Engineer, AI Platform

About Us: Here at Ambience, we never set out to be just another scribe. We’re building the AI intelligence platform that restores humanity to healthcare and drives meaningful ROI for health systems across the country. Our technology helps providers focus on delivering great care by removing the administrative burden that pulls them away from patients and away from their most impactful work. Ambience delivers real-time coding-aware documentation and clinical workflow support across ambulatory, emergency and inpatient settings at the top health systems in North America. Our teams operate relentlessly with extreme ownership to build the best solutions for our health system partners. We value candor, positivity and deep thought — and we expect a lot from each other because we know the problems we’re solving truly matter. Ambience was ranked #1 for Improving the Clinician Experience in the KLAS Research Emerging Solutions Top 20 Report, recognized by Fast Company as one of the Next Big Things in Tech, named one of the best AI companies in healthcare by Inc., and selected as a LinkedIn Top Startup in 2024 and 2025. We’re backed by Oak HC/FT, Andreessen Horowitz (a16z), OpenAI Startup Fund, and Kleiner Perkins — and we’re just getting started. The Role: Ambience ships clinical AI to millions of patient encounters across the nation's largest health systems. How fast we improve that AI depends on the platform you'll own. You'll build evaluation and release gates that let teams ship confidently. Observability that surfaces quality issues before clinicians do. Debug tooling that makes reproducing regressions fast. The chart context retrieval layer that assembles patient history into model-ready inputs. The goal: teams iterate on quality in days, not weeks. Every improvement you make compounds across every product team, every quarter. Our engineering roles are hybrid in our SF office (3x/week). What You’ll Own: - Eval & Release Infrastructure — Automated graders and release gates that work across product pods. Unified eval dataset versioning and execution to replace fragmented workflows. Production quality monitoring with end-to-end tracing, shared metrics, and automated alerting. - Debug Tooling — Encounter replay that reconstructs exact inference inputs (retrieved chart context, packed prompts, model versions) so teams reproduce issues without digging through logs. Diff views comparing known-good runs to regressions. - Chart Context & Data Pipelines — The retrieval layer that pulls relevant patient history and assembles it into consistent model-ready inputs. Feedback loops that capture real-world usage and convert it into training signal. End-to-end latency instrumentation across every workflow step. - Preference Infrastructure — The system that enables clinician and site-specific behavior across specialties. Different clinics want different defaults, different phrasing, different workflows. You'll build the platform that supports customization at scale. - Model Serving — Performance and reliability layer for critical in-house models with clear SLOs, capacity planning, and regression alerts. Who You Are: - 7+ years in software engineering, 3+ focused on ML infrastructure, platform engineering, or data systems - Staff-level scope: owned cross-cutting infrastructure, influenced technical direction across multiple teams - Strong backend fundamentals in Python, TypeScript, or similar - Built eval systems, data pipelines, or ML observability infrastructure - Comfortable on both the ML and Eng sides of MLOps - Track record of platform work that measurably accelerated other teams - In SF, 3x/week in-person Why Here: Healthcare data is messy, customer-specific, and high-stakes. FHIR resources mutate in undocumented ways. Every health system has different mappings. Context windows hit 100K tokens. You're figuring out how to give models the right context for millions of patient encounters across dozens of specialties. Small team, high trust, dire