Event
LLM-Driven Merge Conflict Resolution: From Frustration to Automation
2026-07-28 · 7:00 PM
Online
Recording
About this event
Merge conflicts are an unavoidable part of collaborative software development, often slowing teams down and creating friction during the development process. Recent advances in Large Language Models (LLMs) have opened new possibilities for understanding code context and assisting developers in resolving conflicts more efficiently.
In this talk, Advitya Gemawat, ML Engineer at Microsoft, will explore how LLMs can be applied to merge conflict resolution, the challenges involved in understanding code changes across branches, and the opportunities for building intelligent developer tools.
Drawing from his experience building machine learning and AI systems at Microsoft, Advitya will discuss practical approaches, current limitations, and the future of AI-assisted software engineering.
Whether you're a software engineer, data scientist, researcher, or AI enthusiast, this session will provide valuable insights into how modern AI is transforming the developer experience.
Agenda
- 7:00 PM – Welcome and introductions
- Talk + Q&A
- Wrap-up
Speaker
Advitya Gemawat
Advitya Gemawat is an ML Engineer at Microsoft, specializing in scalable machine learning systems and Responsible AI (RAI). He has authored publications and received awards from leading venues such as VLDB, ACM SIGMOD, and CIDR. At Microsoft, Advitya has worked with Azure Edge & Platform, Gray Systems Lab, and Windows, building ML and LLM services to enhance developer productivity. He also developed Azure ML’s RAI tooling for computer vision models and Azure OpenAI Evaluations, all of which were released at Microsoft Build (2023– 2025). Previously, at VMware, he expanded deep learning features in Apache MADlib. He was a technical reviewer of the Amazon bestseller 'Ace the Data Science Interview' book and was recognized as a “25 under 25: Top Data Science Contributor & Thought Leader.” He is also a keynote speaker at technology panels and podcasts.
Event photos
