Led AI, machine learning, and statistical modeling for a national supply and service operation with $6 billion in annual expenditures across 22 states, ~150 hospitals, and thousands of affiliated care sites. Directed the development of machine learning models (statistical, LSTM, BERT, and LLM-based) to automate product categorization, demand forecasting, and logistics.Enabled scalable analytics by directing the implementation of key technologies including dbt, Prefect, Streamlit, and MLflow, enabling the integration of over 100 data sources in a governed, test-driven, ML-enriched data warehouse. Partnered with non-technical leaders across multiple departments to define success, prioritize roadmaps and delivery plans, and monitor adoption and impact of web applications, dashboards, and other data products.Helped recruit, manage, mentor, and retain an ambitious, creative, and diverse team, and maintained team cohesion and excellence during the transition to remote work.

Co-led analytics and data management for three nine-figure contract events and a new international manufacturing startup.Modeled supply expense using mixed effects models, allowing leaders to monitor performance, trends, and expense recognition issues across care sites.Created hospital-specific profiles of key drivers of performance metrics using boosting and clustering methods.Anomaly detectors caught the first movement of personal protective equipment at the start of the pandemic.Quantified the risk and viability of new clinical programs using simulation and spatial visualization.Standardized data technology and practices, markedly improving efficiency and enabling collaboration across teams, resulting in a promotion to director.

Rescued 3% of annual revenue in two consecutive years for a community mental health network serving ~20k people on ~$100m in annual revenue.Created the company’s first data warehouse, enabling self-service analytics for 800 Tableau users. Advised executives on metric selection and data strategy related to revenue cycle, state contracts, cost accounting, and clinician effectiveness using results from modeling.Used embedding models and NLP to classify the text of police calls, characterizing service utilization across districts.

Modeled risk of psychiatric disorders through the novel application of statistical and machine learning techniques.

Provided evidence-based treatment and assessment in a psychiatric outpatient clinic. Wrote a book chapter.