Arun Iyer

About Me

I am currently a Principal Researcher at Microsoft Research India in Bangalore, where I focus on the intersection of Machine Learning and Artificial Intelligence.

My research journey has been diverse and exciting. I completed my PhD in Computer Science from the Indian Institute of Technology Bombay, working with Prof. Sunita Sarawagi and Prof. Saketha Nath. Before pursuing my PhD, I worked for 4 years as a Research Engineer at Yahoo Labs in Bangalore.

My research interests span a wide range of problems in Machine Learning and AI. I've worked on Neural Networks during my Bachelor's project, Reinforcement Learning in my Master's thesis, Data Mining and Entity Matching problems at Yahoo Labs, and focused on Machine Learning models for estimating label aggregates during my PhD.

Education

PhD in Computer Science

Indian Institute of Technology Bombay (2011-2016)

Thesis: Machine Learning models for estimating label aggregates

M.Tech in Computer Science

Indian Institute of Technology Kanpur (2005-2007)

Thesis: Temporal Difference Learning using Online Support Vector Regression

B.E. in Computer Engineering

University of Mumbai (2001-2005)

Final Year Project: Neural Network Simulator

Research Areas

Artificial Intelligence

Exploring the frontiers of AI to build intelligent systems that can assist and augment human capabilities across various domains.

Data Platforms & Analytics

Developing scalable data processing systems and analytics platforms for extracting insights from large-scale datasets.

Machine Learning & Programming Languages

Bridging the gap between machine learning and programming languages to create more intelligent development tools.

Publications

Here are some of my most cited publications. For a complete list, please visit my Google Scholar profile.

Jigsaw: Large Language Models meet Program Synthesis

Naman Jain, Skanda Vaidyanath, Arun Iyer, Nagarajan Natarajan, Suresh Parthasarathy, Sriram Rajamani, Rahul Sharma

International Conference on Software Engineering (ICSE) 2022

Large pre-trained language models such as GPT-3, Codex, and Google's language model are now capable of generating code from natural language specifications of programmer intent. We view these developments with a mixture of optimism and caution. We present an approach to augment these large language models with post-processing steps based on program analysis and synthesis techniques, that understand the syntax and semantics of programs. Further, we show that such techniques can make use of user feedback and improve with usage.

2022

280 citations

ICSE 2022

CodePlan: Repository-Level Coding using LLMs and Planning

Ramakrishna Bairi, Atharv Sonwane, Aditya Kanade, Vageesh D C, Arun Iyer, Suresh Parthasarathy, Sriram Rajamani, B. Ashok, Shashank Shet

Proceedings of the ACM on Software Engineering, Volume 1, Issue FSE, Article 31

Software engineering activities such as package migration, fixing error reports from static analysis or testing, and adding type annotations or other specifications to a codebase, involve pervasively editing the entire repository of code. We frame repository-level coding as a planning problem and present a task-agnostic, neuro-symbolic framework called CodePlan. CodePlan synthesizes a multi-step chain-of-edits (plan), where each step results in a call to an LLM on a code location with context derived from the entire repository, previous code changes and task-specific instructions.

2024

136 citations

ACM DL

HeteGCN: Heterogeneous Graph Convolutional Networks for Text Classification

Rahul Ragesh, Sundararajan Sellamanickam, Arun Iyer, Ramakrishna Bairi, Vijay Lingam

Proceedings of the 14th ACM International Conference on Web Search and Data Mining (WSDM '21)

We consider the problem of learning efficient and inductive graph convolutional networks for text classification with a large number of examples and features. We propose a heterogeneous graph convolutional network (HeteGCN) modeling approach that unites the best aspects of PTE and TextGCN together. The main idea is to learn feature embeddings and derive document embeddings using a HeteGCN architecture with different graphs used across layers, reducing model parameters significantly and enabling faster training.

2021

114 citations

ACM DL

Maximum Mean Discrepancy for Class Ratio Estimation: Convergence Bounds and Kernel Selection

Arun Iyer, Saketha Nath, Sunita Sarawagi

Proceedings of the 31st International Conference on Machine Learning (ICML 2014), Pages 530-538

We investigate the use of maximum mean discrepancy (MMD) in a reproducing kernel Hilbert space (RKHS) for estimating class ratios in unlabeled collections. Our analysis establishes that the estimate is statistically consistent and provides an upper bound on the error in terms of intuitive geometric quantities like class separation and data spread. We propose a novel convex formulation that automatically learns the kernel to be employed in the MMD-based estimation.

2014

101 citations

PMLR

Active Evaluation of Classifiers on Large Datasets

Namit Katariya, Arun Iyer, Sunita Sarawagi

2012 IEEE 12th International Conference on Data Mining (ICDM), Pages 329-338

We estimate the accuracy of a classifier on a large unlabeled dataset based on a small labeled set and a human labeler. We develop a novel strategy of learning r bit hash functions to preserve similarity in accuracy values and show that our algorithm provides better accuracy estimates than existing methods. Experiments show 15% to 62% relative reduction in error compared to existing approaches, while reading three orders of magnitude less data.

2012

34 citations

IEEE Xplore

Blog

Thoughts on AI, machine learning, and software development practices.

November 13, 2025 AI Tools

📝 Instruction Files for AI Coding Assistants: An Overview

AI coding assistants increasingly rely on instruction files like AGENTS.md, CLAUDE.md, and CODEX.md to guide their behavior. Learn what these files actually do, what belongs inside them, and best practices for using them effectively in your projects.

AI AGENTS.md CLAUDE.md CODEX.md Best Practices

January 14, 2025 AI Tools

🚀 Best Practices for Using Codex CLI and AI Code Generation Tools

Tools like OpenAI's Codex CLI and Anthropic's Claude Code are reshaping how developers work: they can read, modify, and run code locally, acting as an agentic pair-programmer in your terminal. This guide distills practical best practices from official docs, engineering posts, and real-world anecdotes—so you can get reliable, reviewable output rather than "AI spaghetti."

Principal Researcher