Sam J. Miller

• Led a team of 5 data scientists in developing and deploying an anomaly detection system for financial transactions, reducing fraud by 18%.
• Designed and implemented a recommendation engine for a key product, increasing user engagement by 25% and conversion rates by 10%.
• Mentored junior data scientists, overseeing model development, code quality, and productionization.
• Published 3 papers in top-tier machine learning conferences on novel NLP techniques.

• Developed predictive models for customer churn using Python (Scikit-learn, TensorFlow) and large-scale datasets, improving retention efforts by 12%.
• Built and optimized ETL pipelines for various data sources using Apache Spark and SQL.
• Collaborated with product teams to define metrics, analyze A/B test results, and provide data-driven insights for product features.
• Created interactive dashboards using Tableau for key business stakeholders.

Developed a deep learning model (BERT-based) for real-time sentiment analysis on social media data.

• Achieved 92% accuracy on custom datasets for domain-specific sentiment detection.
• Deployed as a microservice using Flask and Docker for easy integration.

Implemented various forecasting models (ARIMA, Prophet, LSTM) to predict product demand with high accuracy.

• Improved inventory management efficiency by 15% through more accurate predictions.
• Built an automated pipeline for model retraining and evaluation.

Ph.D. in Data Science

Carnegie Mellon University, Pittsburgh, PA

Sep 2013 – May 2017

Dissertation: "Optimizing Deep Learning Architectures for Time-Series Prediction"

M.S. in Statistics

University of Michigan, Ann Arbor, MI

Sep 2011 – May 2013