Build a drug access platform for a fast-growing, profitable health-tech company, where your machine learning models directly detect and prevent drug-access failures, helping millions of insured Americans actually receive the medications their benefits promise. We’re hiring a Machine Learning Engineer to work on applied predictive modeling and anomaly detection across complex, high-volume healthcare program data. This is a hands-on, senior IC role focused on turning messy real-world signals into clear, actionable insights that improve medication access across the U.S. healthcare system. Unlike academic or toy ML problems, the work here sits directly in the flow of patient access: identifying failures in drug affordability programs, surfacing non-obvious breakdowns in benefit usage, and enabling teams to intervene before patients abandon treatment. What You’ll Do - Build predictive models (forecasting, time series, trend analysis) on real-world program and financial data. - Design anomaly / outlier detection to surface risks, inefficiencies, and non-obvious patterns. - Own feature engineering, model evaluation, and interpretability (feature importance, SHAP-style analysis). - Work end-to-end from exploration → modeling → business-facing insights. - Partner with product and analytics teams to turn models into decisions. Must-Have - 4–7 years as ML Engineer or Applied Data Scientist - Strong predictive modeling + anomaly detection background - Hands-on experience with messy, high-volume financial or program data - Strong statistical intuition (not just fitting models) - Python with Pandas / NumPy / scikit-learn - Comfortable operating independently as a senior IC Nice to Have - U.S. healthcare-adjacent data (insurance, benefits, copay, utilization) - Experience in compliance-aware environments (HIPAA-constrained data) - Translating models into dashboards, reports, or exec-ready insights - XGBoost / LightGBM, solid SQL Explicitly Not Required - MLOps ownership or DevOps work - Managing teams - Pharma biology expertise \n \n