Leading efforts on generative AI and agentic “guardrails” ensuring that unsuitable prompts or unwanted output of LLMs and agents will be limited in watsonx and Red Hat Openshift AI applications

Maintained open-source projects including caikit - https://github.com/caikit

Architected and developed infrastructure to provide trust algorithms and metrics, including explainability, fairness, and uncertainty quantification algorithms

Led efforts to design, develop, and optimize infrastructure for the training and deployment of NLP models, allowing product users to use models to analyze text features