I build ML systems that change how products work and how decisions get made.
Staff Data Scientist at Adobe. Previously Accenture, Stanford, Caltech.
Next-best-action models that drove 20% lift in feature adoption across thousands of users
Measuring what works: A/B tests, natural experiments, and scalable experimentation platforms
Early warning systems that flag $16M+ in at-risk revenue for proactive intervention
Refugee placement optimization presented to governments in 4 countries
Synthetic controls show Twitter's content moderation reduced election misinfo flow by ~15%
ML for tracking fast-moving online conversations—37% better at catching trends than static methods
Diff-in-diff on Twitter data shows Supreme Court rulings polarize opinion in affected states
Word embeddings to auto-discover keywords for detecting online harassment
ML recommendations for where immigrants should settle—simulations show higher earnings