About me

I am an Assistant Professor in the Industrial Engineering Department at the University of Pittsburgh. Before joining the University of Pittsburgh, I was a visiting postdoctoral researcher at Google in the Optimization and Algorithms group in New York. In 2019, I received my Ph.D. from the Department of Management Science and Engineering at Stanford University where Professor Yinyu Ye was my advisor.


My research focuses on continuous optimization, with a penchant for local optimization methods such as gradient descent. I aim to develop reliable and efficient algorithms built on solid mathematical foundations. The goal of my research is to build new optimization tools for operations research and machine learning. I typically aim to improve existing optimization software in terms of speed, scalability, accuracy, usability, reliability and representability. My work spans the spectrum from fundamental optimization theory to software development. I build both general purpose optimization solvers and solvers targeted at specific applications such as drinking water networks, electric grids, and certifiable deep learning.

You can contact me at ohinder at pitt dot edu.

Please see my scholar page for an up to date publications list.

My Ph.D. thesis was Principled Algorithms for Finding Local Minimizers.


Industry impact

My research has had direct impact on industry: my research on nonconvex IPM inspired LocalSolver's own implementation, my work on first-order methods for linear programming is soon to be part of the Google Operation Research tools package, and my work with DeepMind has led to state-of-the-art techniques for neural network certification.

Papers and talks

Grouped by topic

First order methods for convex optimization

Recorded talk: https://www.youtube.com/watch?v=aViqFWsrT2M

Structured nonconvex optimization

The complexity of finding stationary points of nonconvex functions

Slides from my 2019 ICCOPT talk summarizing this body of work.

Machine scheduling and integer programming