DUET: Co-Evolving Solver and Grader Agents ↗
Fengyu Gao, Sourav Pal, Austin Z. Henley, Arjun Radhakrishna, Gustavo Soares
Jointly improving agents that solve tasks and evaluate their outcomes.

I work at Google. My interests are in AI, software engineering, and human-computer interaction, with a focus on agents, realistic benchmarks, and learning from feedback.
I joined Google in October 2026 after nine years at Microsoft, where I worked on program synthesis tools for Visual Studio, early Copilot Chat and agent experiences, and most recently agents for Excel. I received my Ph.D. from the Federal University of Campina Grande in 2014 and was a postdoctoral researcher at UC Berkeley.
Fengyu Gao, Sourav Pal, Austin Z. Henley, Arjun Radhakrishna, Gustavo Soares
Jointly improving agents that solve tasks and evaluate their outcomes.
Aayush Kumar, Avik Dutta, Sumit Gulwani, Gustavo Soares, Advait Sarkar, Emerson Murphy-Hill
How upfront planning changes the experience of working with spreadsheet agents.
Avik Dutta, Priyanshu Gupta, Hosein Hasanbeig, Rahul Pratap Singh, Harshit Nigam, Sumit Gulwani, Arjun Radhakrishna, Gustavo Soares, Ashish Tiwari
Evaluating conversational data analysis when goals need clarification.
Shashank Kirtania, Param Biyani, Priyanshu Gupta, Yasharth Bajpai, Roshni Iyer, Sumit Gulwani, Gustavo Soares
Turning experience into structured, reusable knowledge. COLM 2026.
Aayush Kumar, Yasharth Bajpai, Sumit Gulwani, Gustavo Soares, Emerson Murphy-Hill
Studying the role of developer involvement in agent-assisted software engineering.
Jenny T. Liang, Aayush Kumar, Yasharth Bajpai, Sumit Gulwani, Vu Le, Chris Parnin, Arjun Radhakrishna, Ashish Tiwari, Emerson Murphy-Hill, Gustavo Soares
Supporting spreadsheet development through conversational, incremental interaction.