Using AI Agents To Test GenAI Software
12:45 - 1:45 Student Alumni Room
As Generative AI reshapes software development, ensuring the quality and reliability of GenAI-generated code presents significant challenges for development teams — especially those with limited resources. Traditional testing workflows often fall short in identifying subtle bugs, maintaining compliance, and addressing corner cases in complex codebases. This session explores how autonomous AI testing agents have evolved beyond co-pilot assistance to full autopilot capabilities, enabling seamless, end-to-end quality assurance with minimal human intervention.
What Attendees Will Learn:
* The differences between AI co-pilot and autopilot testing solutions and their impact on testing workflows.
* Practical steps to integrate autonomous AI testing tools into DevOps pipelines with minimal disruption.
* Strategies for achieving robust code integrity in GenAI-driven environments while addressing common challenges like missed corner cases and compliance risks.
* Key metrics and KPIs for evaluating the success of AI-powered testing solutions in resource-limited teams.
Yunhao Jiao, a Yale University graduate with a master’s degree. He has nearly five years of experience at AWS, where he served as a Senior Software Engineer. During his time at Amazon, he played a crucial role in building the Contract Tests framework for AWS CloudFormation, ensuring that resource types operated as expected throughout their lifecycle. Since 2015, Yunhao has been actively involved in NLP research, publishing his first AI paper as the lead author in 2017. He also contributed to education as the Chinese High School Artificial Intelligence Textbook editor. After his tenure at Amazon, Yunhao founded TestSprite and participates in the Techstars and YC China accelerator programs.