Portrait of Arjun Radhakrishna

Arjun Radhakrishna

Principal Researcher

Connect on LinkedIn

About

I am Researcher in the PROSE team (opens in new tab). My  research is motivated by the question “Can computers help me build better systems quicker?” I focus on techniques based on program synthesis and formal methods with applications geared towards programmers, students, and data scientists. I have worked on synthesis of concurrent programs, quantitative synthesis from soft specifications, and synthesis of automated code edits.

Recently,  I have been exploring the idea of modeless synthesis: here, the user does not explicitly provide a specification for synthesis, either in terms of examples or functional specifications. Instead, the user goes about their work as usual, and the specification is automatically inferred. Users are more likely to use program synthesis based techniques when they do not have to break their regular workflow to provide examples or specifications. In this line, we have developed a system for automated suggestions for refactorings and repetitive code edits that has been released as the IntelliCode Suggestions (opens in new tab) features in Microsoft Visual Studio 2019. Here are two videos demonstrating the feature:

Previously, I was a post-doctoral researcher at the University of Pennsylvania working in the group of Prof. Rajeev Alur (opens in new tab). Before that, I completed my PhD at IST Austria (opens in new tab) under the guidance of Prof. Thomas A. Henzinger (opens in new tab). Check out my dissertation here (opens in new tab) on Quantitative Specifications for Verification and Synthesis. My dissertation was awarded the 2014 ACM SIGBED Paul Caspi Memorial Dissertation Award.