data
Posted 22 hours agoData Scientist
at Metriport
San Francisco, United StatesOn-site
Requirements
- ABOUT YOU We're looking for a data scientist who thrives in ambiguity and cares as much about getting the fundamentals right as building the next model.
- If a metric is off by 1%, it keeps you up at night until you find the root cause. - You believe that high-quality clinical data is the bedrock of excellent healthcare, and you're excited to work at the intersection of ML and patient records. - You have a strong sense of ownership and the ability to lead cross-functional initiatives with minimal direction. - You care about impact over sophistication.
- Day to day, this looks like: - Applying AI/ML to Clinical Data: Building and deploying models to predict patient outcomes, identify gaps in care, or surface anomalies across our data warehouse. - Normalizing Clinical Data at Scale: Using NLP, LLMs, or rule-based systems to transform messy, unstructured clinical records into structured, searchable, trustworthy data. - Owning Analytics When It Matters: You'll share ownership of our analytics stack and data quality alongside the team.
- experience in a data science role, ideally at a high-growth or startup company where you wore multiple hats. - SQL mastery: You can write complex, performant queries in your sleep. - ML/statistical modeling: Practical
- experience building and deploying models (classification, regression, clustering, NLP) — not just prototyping. - Coding proficiency: Strong in Python (pandas, scikit-learn, and ideally some
- experience with LLM APIs or frameworks).
- TypeScript proficiency is a plus — our stack is TypeScript-heavy. - Analytical chops: You're comfortable owning dashboards, data quality, and ad-hoc analysis.
- Experience with healthcare data is strongly preferred — FHIR, HL7, or clinical data. Understanding how a patient moves through the healthcare system is the core of what we do. -
- Experience with data modeling tools (dbt or similar) and product analytics platforms (PostHog, Mixpanel, Amplitude). -
- We use dbt for transformations and Posthog for product analytics.
- Our infrastructure is managed via AWS CDK, and our core platform is written in TypeScript and Python.
Experience
- REQUIREMENTS - 4+ years of