Keynote du LIG - Behrooz Omidvar-Tehrani, Shweta Garg

Abstract

As LLM agents are evolving from passive assistants to proactive collaborators, we are witnessing a transformation in how humans and machines work together. This talk explores the foundations and future of human-AI collaboration, tracing its roots from Licklider’s 1960 vision of man-computer symbiosis to today’s emerging generation of AI teammates. Drawing on research and real-world deployments, we examine the modalities of human-AI interaction and argue that the most effective systems keep humans in the loop for oversight, creativity, and judgment. We will explore two major domains where AI teammates are making an impact: software development and social collaboration. In development, coding agents are evolving from autocomplete tools to autonomous collaborators that plan, execute, and adapt alongside developers. In social and knowledge workspaces, collaborative agents are beginning to support communication, coordination, and content creation, often operating in the background to enhance productivity and connection. Through case studies like Amazon Q and tools like GitHub Copilot and Amazon Kiro, we will examine how AI teammates are reshaping workflows, enabling parallelization, and introducing new forms of specialization. Finally, we will explore the architectural and cognitive foundations of these agents, including planning, memory, and multi-agent collaboration, and reflect on the long-term vision of AI teammates that augment human potential.

Biography

Behrooz Omidvar-Tehrani is a Principal Applied Scientist at Microsoft, where he advances agentic AI systems for M365 Copilot to improve reasoning and engagement across communication-centric workflows. Previously, he was a Senior Applied Scientist and Science Manager at Amazon contributing to Amazon Q Developer and AWS Transform. He earned his PhD in 2015 from the University of Grenoble Alps in France. He has published more than 40 papers in top-tier conferences including ICML, ICLR, SIGMOD, CHI, and CIKM.

Shweta Garg is a Senior Applied Scientist and Applied Science Manager at AWS, where she leads the development of Q Agentic Chat for the Command Line and IDE Plugins. With 12+ years of experience in AI, her current work focuses on building collaborative, agentic coding assistants that leverage memory, planning, tools, and project context to enable natural language-driven software development. She drives end-to-end innovation across model development, product integration, and rigorous evaluation using both online metrics and offline benchmarking. Her research spans NLP, generative AI, and code intelligence, and she has published at top-tier venues including EMNLP, ECNLP, and AMLC. She has also organized ML workshops at AMLC and delivered tutorials at Amazon’s ML School.