I am a Principal Research Staff Member and Manager of the Cognitive Systems Performance Lab in the Cognitive Computing Research Organization.
The Cognitive Systems Performance Lab (CSPL) mission is to respond to the challenges that arise when deploying cognitive capabilities in real-world production settings. Theses challenges include harvesting, boosting, and maintaining high-quality ground truth data that is needed to train cognitive services. Once in production, cognitive services are continuously monitored and CSPL focuses on developing adequate metrics and tooling for performance analysis, diagnostics and repair to drive continuous improvements. Another important goal of CSPL is to significantly reduce time-to-value through new tooling and methodologies for fast domain adaption and active learning to support bootstrapping as well as the ongoing evolution of cognitive services.
In addition to these core CSPL themes, and in collaboration with our security research organization, we are working on emerging directions in the area of resilient and trusted cognitive computing. Cognitive APIs face unique security threats that aim to corrupt or mislead machine learning models through seemingly legitimate use of an open API access point. We are working on a comprehensive framework for data, model and operations security to protect and assure cognitive APIs.
In the past I have also worked in the areas of parallel program analysis, dynamic and static program optimization, and software development tools, including cloud-based development tools.
I am an ACM Distinguished Scientist.