Co-founder & CEO at PrimaLabs.
Ex-Director of AI at ORNL.
I bridge the gap between Exascale research and commercial AI. With a background managing a $200M+ portfolio and training trillion-parameter models on the world's fastest supercomputers, I now build secure, energy-efficient AI systems that solve hard problems.
Directed the first trillion-parameter AI model training on the Frontier supercomputer (world's fastest). Proved capability to scale foundation models beyond industry standard limits.
Developed ORBIT, a 113-billion parameter Vision Transformer. Surpassed previous climate AI foundation models by 1,000x in size, earning a finalist spot for the ACM Gordon Bell Prize.
Secured $20M+ in direct funding and managed an $8M LDRD portfolio. Established R&D partnerships with NVIDIA, OpenAI, Microsoft, and Google.
Creator of DeepHyper (scalable AutoML) and ytopt. My work in automating machine learning workflows is used to optimize complex scientific systems efficiently.