Summary of research interests:

  • high-dimensional data analysis
  • time series analysis
  • statistical/machine learning
  • computational statistics
  • graphical models

My research interests are at the intersection of applied, computational and theoretical statistics. Being statistician with BS in Computer Science and Master in Economics, I enjoy solving applied problems which lead to improving existing statistical methodologies. I believe every high-dimensional data has underlying, scientifically meaningful, low-dimensional structures and one of my primary goals is to assist in identifying and discovering this structure. I am convinced that in nowadays fast growing world, collaboration with within and outside field scientists is essential in achieving this goal.

The methodologies I am interested in are in the areas of high dimensional time series analysis, machine learning, and optimization. The classical covariance estimation in multivariate and time series analysis perform poorly when applied to modern datasets due to the presence of more variables than observations. I am interested in addressing these problems jointly in the high-dimensional settings by developing a new statistical methodology that is both computationally efficient and theoretically sound. I have found that penalization techniques and the tools of convex optimization are particularly useful in achieving this goal.

I have worked on a variety of applied problems such as discovering the underlying structure in food recall, high dimensional financial datasets, and call center data. My methodological work has been strongly motivated by these applied projects and I am looking forward to new collaborations in the future.