Location: REMOTE / Nashville, Tennessee
This job allows you to work remotely.
Our client is transforming the experience of specialty care. Their comprehensive care program takes a profoundly personal, evidence-based approach to improving patient outcomes for joint, back, and muscle conditions. By carefully assessing patients’ symptoms, health histories, preferences, and goals with predictive data and the latest evidence-based guidelines, they help patients choose and navigate the most effective treatment pathway every step of the way.
The company values the experiences and perspectives of individuals from all backgrounds. They are a highly collaborative, curious, and determined team passionate about scaling a high-growth start-up to improve the lives of those in pain.
The Role: Director, AI - Decision Intelligence
You will lead the team that turns our organization's data into decisions: who we reach, how we target outreach, which care pathway we recommend, and how we forecast clinical and financial outcomes. This is the ML and decisioning core of the company. The models your team ships directly drive patient engagement, procedure avoidance, and partner savings.
We are growing fast. We are adding payers and markets quickly, and the systems and team you own have to scale with that pace. We need a leader who can deliver against near-term launch commitments while building for an order of magnitude more volume, grow and level a team through that change, and stay effective when priorities shift underneath them. Comfort with ambiguity and a bias toward execution matter as much as technical depth here.
This is a player-coach leadership role. Our teams own and drive outcomes, not task lists. You will be accountable for results, with the latitude and the obligation to decide how your team gets there. You will own the team strategy and delivery, set the technical bar, and stay close enough to the work to make architecture and modeling calls yourself.
You will:
Lead a team of outcome-driven data scientists and ML engineers, with direct accountability for delivery, technical quality, and growth.
Drive cross-functional partnership with Medical Economics, Clinical Operations, Product, and the data platform team.
Own the interface between modeling work and the platform and infrastructure it runs on.
Make build-versus-buy and architecture calls, set the technical bar, and stay hands-on enough to make modeling decisions yourself.
Must Have Skills:
Master's or PhD in a quantitative field (computer science, statistics, machine learning, operations research, applied mathematics, economics, or a closely related discipline). This is a requirement for the role; a PhD with applied, production-oriented research is a strong plus.
A demonstrable history of ML systems you shipped to production that moved a business or clinical metric, with the specifics of what you built, what changed, and how it was measured.
Evidence of delivering against hard external deadlines and managing data-dependency risk without slipping quality.
A record of building and growing high-performing technical teams, including hiring, leveling, and developing data scientists and ML engineers.
Experience owning a model portfolio across its full lifecycle, retiring or refactoring models that no longer earn their place.
Preferred qualifications:
Healthcare, payer, or value-based care experience, and familiarity with HIPAA-regulated data.
Experience translating actuarial or medical-economics concepts into model features and targets.
Published or peer-reviewed work in applied ML, forecasting, or causal inference.
What you Bring:
Deep applied ML: supervised learning on tabular and structured data, gradient-boosted trees (XGBoost, LightGBM), feature engineering, calibration, and rigorous offline and online evaluation.
Production ML engineering: model packaging, deployment, monitoring, drift detection, and retraining pipelines. You own model quality in production, not just in a notebook.
Strong software engineering fundamentals: Python, SQL, version control, testing, and code review standards you can set and enforce.
Modern data and ML platform fluency: Databricks, dbt, and AWS (S3, Postgres, DynamoDB). Comfortable making build-versus-buy and architecture calls.
Experimentation and causal rigor: A/B testing, uplift modeling, and the judgment to distinguish correlation from decision-relevant signals.
Sound judgment on where newer methods (LLMs, agents, feature augmentation from external signals) add measured lift versus where they add cost and risk.
You lead with the recommendation and state risks plainly, escalate risk early, and decide fast. Communication is concise and structured.
