Hello! I'm a postdoc in the Empirical Inference Department at MPI. Until recently, I was a software engineer at Google (Pittsburgh), and, before that, I did my PhD in Machine Learning and Statistics at Carnegie Mellon University, advised by Barnabás Póczos.

My main research is in statistical machine learning - using statistical models to understand when and why machine learning algorithms work. I also develop models and tools (of math, stats, and machine learning varieties) to help researchers in biology, psychology, and other sciences analyze data.

Some specific things I've worked on, by topic:

Statistical Machine Learning

• Statistical guarantees (minimax bounds) for:
  • Generative Adversarial Networks (GANs) and other density estimators
  • Estimating entropy/mutual information/divergence, smoothness, and other properties of probability distributions
  • Multi-class/imbalanced classification
  • Unsupervised learning of shallow convolutional neural networks
  • Zeroth-Order/Blackbox/Gradient-Free Optimization

Applied Machine Learning

• Forecasting the effects of the COVID-19 epidemic (at Google)
• Compressing and accelerating deep neural networks (at Amazon)

Computational Biology

• Deep learning to predict gene interactions from DNA sequence
• Differential equations and multi-agent simulations to model bacterial swarm chemotaxis

Psychology

• Probabilistic modeling and deep learning to analyze eye-tracking data and study sustained attention development in children