Led ideation and development of novel ML models by identifying high-value prediction problems from existing data assets. Designed statistical experiments incorporating causal inference techniques (difference-in-differences, propensity matching) to validate feasibility and improve predictive accuracy. Coordinated with cross-functional teams to assess data source availability and quality, streamlining the ideation-to-prototype workflow. Built prototype models adhering to MLOps standards with versioned repositories and documented training pipelines, enabling smooth handoff to the production team. Authored model cards and pre-registrations detailing problem statements, feature definitions, and known failure modes to ensure transparency across the model lifecycle. Leveraged AI-assisted coding tools to accelerate exploratory analysis and training, while maintaining high-quality code. Performed power analyses to size experiments and validate assumptions prior to prototyping. Presented finding

Data Scientist (ML & Analytics) at Databricks

December 1, 2021 - July 1, 2023

Developed and validated machine learning models (supervised learning) addressing diverse business challenges, improving decision-making processes across departments. Performed feature engineering and exploratory data analysis to identify key predictors and boost model performance, contributing to a 30% increase in prediction accuracy. Collaborated with data engineering to ensure data quality and governance (HIPAA/PII considerations). Designed controlled experiments with baseline benchmarks to validate model viability before deployment. Authored model cards and training pipelines to support reproducible handoffs during production. Drove proactive model ideation, resulting in two new ML models reaching production state. Used cloud ML platforms (Azure ML, Databricks) for scalable training and deployment. Communicated findings to stakeholders to drive data-driven decisions and collaborated with cross-functional teams to align initiatives with business objectives. Mentored interns and junio