Gabriel Ryan

I am a Senior Research Scientist at Microsoft, where I work in the CodeAI organization training specialized models behind GitHub Copilot. My work spans the underlying models that power AI-driven software development — code completion, context retrieval, agentic coding workflows, and automated code review.

Current Work at Microsoft CodeAI

Copilot Subagents & Context Management

Training custom models for Copilot subagents, context compaction, and memory generation — enabling agentic workflows to delegate focused sub-tasks and operate over long horizons.

Context Retrieval Embeddings

Trained the embedding models used to index GitHub for Copilot's context retrieval. [GitHub blog]

Code Completion Models

Trained the code completion models that power GitHub Copilot in editors used by millions of developers.

Copilot Code Review

Trained code quality evaluation models used in Copilot Code Review, helping the system identify substantive issues in pull request changes.

Background

I received my PhD from Columbia University in 2023, advised by Prof. Suman Jana, where I worked on AI-driven security and software development. During my PhD I interned at AWS AI Labs (2023) and Microsoft Research RiSE (2021), resulting in publications at FSE ‘24 and ICSE ‘22 (Distinguished Paper Award).

Selected Publications

ICML · 2026

SemRep: Generative Code Representation Learning with Code Transformations

W. Li, J. Song, B. A. Stoica, A. Dhoot, Gabriel Ryan, S. Fu, K. Pei

Generative learning of code representations using semantics-preserving code transformations as a self-supervised signal.

paper

NeurIPS DL4C Workshop · 2025

DevBench: A Realistic, Developer-Informed Benchmark for Code Generation Models

P. Ameneh Golnari, A. Kumarappan, W. Wen, X. Liu, Gabriel Ryan, Y. Sun, S. Fu, et al.

A realistic, developer-informed benchmark for evaluating code generation models on tasks that reflect real-world software development.

paper

FSE · 2024

Code-Aware Prompting: A study of Coverage Guided Test Generation in Regression Setting using LLM

Gabriel Ryan, Siddhartha Jain, Mingyue Shang, Shiqi Wang, Xiaofei Ma, Murali Krishna Ramanathan, Baishakhi Ray

Static-analysis-guided prompting that drives LLMs to reason symbolically about execution paths, doubling coverage of generated regression tests.

paper

ICSE · 2022 (ACM Sigsoft Distinguished Paper Award)

TOGA: A Neural Method for Test Oracle Generation

Elizabeth Dinella*, Gabriel Ryan*, Todd Mytokowitz, Shuvendu Lahiri

A neural transformer paired with a test-oracle grammar that automatically generates bug-finding assertions, with a 170% improvement in bug detection over prior systems.

paper code

OSDI · 2021 (OSDI Jay Lepreau Best Paper Award)

DistAI: Data-Driven Automated Invariant Learning for Distributed Protocols

Jianan Yao, Runzhou Tao, Ronghui Gu, Jason Nieh, Suman Jana, Gabriel Ryan

Data-driven inference of inductive invariants for verifying safety properties of distributed protocols.

paper code

See the full list of publications for all papers.