Anand Ashok Sawant

Postdoctoral scholar at University of California, Davis

I am a Postdoc working in the Decal lab at UC Davis with Prof. Premkumar Devanbu. My current research focus is on applying natural language models to new software engineering tasks. My latest work has centered around improving the understandability of decompiled binary code using a pre-trained language model for C. During my PhD I focused on the human aspect of API evolution to understand how and why APIs evolve and its impact on API consumers. Going forward I would like to enhance the API evolution process by devising a semantic aware approach that is able to automatically migrate consumer code from one API to another.

In my spare time I enjoy taking in a lot of sports (favorite sport teams include: Manchester United, Indian cricket team and LA Lakers), hiking in the hills and traveling to far flung places of the world. Fun fact: the background image of this page is generally related to the area I am currently in.

Education

University of California davis

Postdoctoral scholar
Machine Learning for Software Engineering

Advisor: Prof. Premkumar Devanbu

October 2019 - Ongoing

Delft University of Technology

PhD candidate
Software Engineering

Advisor: Alberto Bacchelli
Promotor: Arie van Deursen
Thesis title: The impact of API evolution on API consumers and how this can be affected by API producers and language designers

November 2015 - October 2019

Delft University of Technology

Masters of Science
Computer Science - Software Technology Track

Thesis grade: 9.0
American equivalent: A+
Thesis advisor: Alberto Bacchelli
Thesis title: fine-GRAPE: fine-Grained APi usage Extractor An Approach and Dataset to Investigate API Usage

August 2013 - October 2015

Delft University of Technology

Bachelors of Science

Thesis grade: 9.0
American equivalent: A+
Thesis advisor: Alexandru Iosup
Graduated with honors and cum laude

August 2010 - July 2013

Peer Reviewed Publications

Research Interests

I am a Postdoctoral scholar working on the application of Machine Learning for Software Engineering at the Decal lab at the University of California Davis, USA. My current research focuses on improving data collection for deep learning tasks. My work on improving data fidelity makes tasks such as the recovery of inlined functions from binaries and variable names and types from obfuscated code achievable. Furthermore, I seek to enhance current deep learning models for code with a notion of semantic understanding by leveraging execution traces generated on a large scale, thus making complex tasks such as automatically dealing with API evolution realizable. My earlier research centered around mining API usage and understanding API evolution by conducting empirical studies on API usage data and conducting qualitative studies to understand the developer perspective on API evolution.

Contact Me