Data Scientist | Individual Contributor | Data & Machine Learning

Description

Join the team shaping the future of healthcare! Medbridge is a dynamic software as a service company working with the country’s largest healthcare providers to build technology solutions helping patients get better faster, while decreasing the overall cost of care.

• 15 years serving Acute, Post-Acute, and Ambulatory settings.
• 4,000+ organizations, 360,000+ healthcare providers, and 15 million patients served annually.
• 9 of the top 10 largest hospitals and health systems, 9 of the top 10 largest private practices, and 6 of the top 10 largest home health agencies run on Medbridge.

• We are mission driven. Our mission is clear, we want to help everyone move well, feel well, and live well.
• Medbridge's Educate is built in-house by 600+ named instructors, many of whom wrote the guidelines clinicians already follow.
• Medbridge's Care gives those clinicians home exercise programs, patient education, clinical pathways, remote therapeutic monitoring, and outcomes collection inside the workflows they already use
• You’ll shape the Data Science practice. This is Medbridge's first dedicated data science role, reporting to a Head of Data & ML who is a statistician by training and still writes code. You will set the modeling standards the team follows.

We are a remote first company and we hire ONLY in the following states: AZ, CO, FL, GA, , IL, KS, MA, MI, MN, MO, NH, NY, NC, OH, OR, PA, SC, TN, TX, UT, VA, WA, and WI.

The Staff Data Scientist is the senior individual contributor responsible for Medbridge's predictive and statistical models. The core problems are hard and consequential: comparing patient outcomes fairly across providers when patients differ, predicting what is likely to happen next, estimating whether a product or program changed anything, and, over time, recommending what to do. You own each problem end to end: framing it with product and clinical partners, choosing the method, building and validating the model, shipping it to production in Snowflake, monitoring and retraining it, and measuring whether it is used and what it changed.

The role leads through influence rather than direct people management.

• Own the design, development, validation, and production operation of Medbridge's predictive and risk-adjustment models, beginning with the models behind our outcomes benchmarking product.
• Set the methodological bar: choose approaches that fit the question and the data, validate to a standard that holds up to outside expert scrutiny, and document decisions so partners understand and can question them.
• Write production-quality code: tested, versioned, reviewed, and reproducible, and own its deployment and monitoring in Snowflake in partnership with Data Engineering.
• Estimate whether Medbridge products, programs, and content change patient outcomes, using experimental and observational methods, and shape the instrumentation that makes those estimates possible.
• Turn ambiguous goals into well-posed problems with options and tradeoffs, work directly with product, engineering, clinical, and customer-facing teams to get models into the product, and establish the team's practices for evaluation, documentation, monitoring, and responsible use of healthcare data.

• A track record of models that ran in production and held up: you have owned model code through deployment, monitoring, and retraining, and you can describe what broke and how you fixed it.
• Statistical depth you use every day: you reason about how the data were generated, recognize when a model's assumptions fail, quantify uncertainty, and can explain why one method fits a problem and another does not.
• Strong Python and SQL and disciplined software engineering, including testing, version control, code review, and reproducible pipelines. Snowflake experience is strongly preferred.
• Experience with messy real-world healthcare data, including product data used by clinicians, administrators, or patients, and working knowledge of causal inference and evaluation design.
• Experience taking a model from an open-ended business goal to a shipped product feature in partnership with product and engineering, and comfort being the first person on a problem.
• Judgement under constraint: when the data, timeline, or requirements are not what the ideal method assumes, you find a path to a defensible solution, explain the tradeoffs plainly, and know what would let you strengthen it later.
• Clear…