Getting to know me better

Hi there, I’m Alex Mathai! I am a computer science PhD student at Columbia University. My PhD research combines AI and software engineering – where I design LLM Agents to resolve bugs in million-line codebases.

Before this, I was a Research Engineer at IBM Research. At IBM, I integrated practical AI tools for open-sourced and commercial software. My previous contributions include modernizing monolithic applications using AI, classifying & analyzing internet traffic at gigabyte scales, and making AI models robust in real-life scenarios.

Where am I now ?

I am currently a PhD student at Columbia University, New York. I am grateful to be advised by Professor Baishakhi Ray, Professor Junfeng Yang, and Professor Suman Jana

I currently focus my time and effort on designing LLM Agents that help assist in the bug resolution process of vast, complicated, and challenging repositories like the Linux kernel (20 million lines long and counting). 

 

Research Interests

ML

Applications

LLM Agents

ML for Software Engineering

Deep

Learning

Transformers

Graph Neural Networks

Unsupervised

Learning

Clustering

Auto Encoders-Decoders

Industrial & Academic Experiences

IBM Research AI & Cloud, Research Engineer

From 2020-23, I was a research engineer at IBM Research, where I worked on multiple projects at the intersection of AI and software engineering. Collaborating with Amith Singhee (Director, IBM Research), I modeled large monolithic codebases as knowledge graphs. I then partnered with Srikanth Tamilselvam (Research Scientist) to apply graph neural network techniques that partition these graphs and suggest candidate microservices.

Canopus Networks, Intern

I worked with Prof Vijay Sivaraman and Prof Hassan Habibi at the University of New South Wales on the classification of encrypted network traffic. My contributions helped provide valuable insights on internet traffic composition. I also created statistical models that could accurately estimate user experience for live streaming videos. The deep learning models I trained were deployed in 8000 Australian homes.

IBM Research AI, Intern

I interned under Shreya Khare, Srikanth Tamilselvam and  Senthil Mani (Senior Technical & Senior Manager) at IBM Research. My work primarily involved creating adversarial attacks for natural language models and inferring their robustness concerns. I was successful at creating adversaries for many popular text models by combining genetic algorithms with deep learning models.

Indian Space Research Organisation (ISRO), Intern

I interned at the Indian Institute of Remote Sensing (ISRO Dehradun) where I worked with Sameer Saran (Geoinformatics Department Head) and Ashutosh Kumar Jha (Research Scientist). I worked on creating a customized U-Net for Satellite Image Feature Detection in High-Resolution Multispectral Satellite Images.

Central Electronics Engineering Research Institute, Intern

I was part of a project that developed AI-powered Surveillance Systems. Under the guidance of Sanjay Singh (Research Scientist) and Sumeet Saurav (Research Scientist), I successfully implemented an Unsupervised Deep Learning Model using Generative Adversarial Networks for Optical flow prediction.

Close Menu