Prasanna Balaprakash
Co-founder & CEO at PrimaLabs, building next-generation hardware, application, cost -aware enterprise AI solutions. Former Director of AI and Distinguished R&D Scientist at Oak Ridge National Laboratory and Group Leader at Argonne National Laboratory, where I directed breakthrough research at the intersection of AI and high-performance computing.
Track Record
Two decades of experience leading AI research and development at the world's most advanced computing facilities. I bridge fundamental research and real-world deployment, from training trillion-parameter models on exascale supercomputers to building optimization platforms used across government, industry, and academia. Led research efforts at Oak Ridge in launching transformational mode consortium (ModCon), which is a basis for DOE Genesis Mission to accelerate science through AI.
Breakthrough Projects
Exascale AI Training
Directed the first trillion-parameter AI model training on Frontier, the world's fastest supercomputer. Achieved 87% GPU efficiency on AMD MI250X hardware—proving capability to scale foundation models beyond industry standard limits across 49,152 GPUs.
ORBIT Foundation Model
Directed the development of a 113-billion parameter Vision Transformer for climate modeling. Surpassed previous climate AI models by 1,000x in scale, achieving 1.6 exaFLOPS sustained throughput. ACM Gordon Bell Prize Finalist (2024, 2025).
DeepHyper Platform
Created DeepHyper, a scalable AutoML framework for neural architecture and hyperparameter search. Powers AI optimization across national laboratories, enabling 10x faster deployment on new hardware platforms.
Graph Foundation Models
Pioneered large-scale graph foundation models for atomistic materials modeling on 5.2TB of materials data. Achieved near-linear scaling using 16,000+ GPUs, advancing AI-driven materials discovery.
Hardware Co-Optimization
Developed ytopt, a hardware-aware optimization toolkit that automates performance tuning across diverse computing architectures.
AI for Climate Science
Built hybrid climate modeling systems using AI-assisted process emulators. Demonstrated long-range weather prediction capabilities that bring operational forecasting within reach for extreme weather events.
Industry Collaboration
Established strategic partnerships that accelerated AI innovation and provided early access to next-generation platforms.
Strategic Impact
ORNL AI Initiative Director
Led ORNL's $8M/year AI Initiative (lab's largest cross-cut LDRD investment), managing 50+ researchers across 10+ advanced AI projects. Restructured the initiative to prioritize secure, trustworthy, and energy-efficient AI for scientific discovery and national security.
Section Head, Data & AI Systems
Provided strategic leadership for 45 researchers across five groups specializing in AI/ML, visualization, data workflows, and performance systems. Shaped divisional research directions through team building and resource alignment.
Policy & Advisory
Member of Tennessee AI Advisory Council (Governor-appointed). Invited participant at White House roundtables on AI R&D. Delivered briefings to U.S. Senate and House staffers on AI strategy and supercomputing.
Program Development
Obtained over $20M+ in direct funding and helped secure over $200M+ in total funding across 44 competitively reviewed proposals (33 external, 11 internal). This substantial financial support has played a pivotal role in advancing research and development across AI, machine learning, high-performance computing, and scientific applications.
Awards & Fellowships
Research & Dissemination
Two decades of contributions spanning AI/ML systems, high-performance computing, and scientific applications. Research published in top-tier venues including SC, NeurIPS, ICML, Nature journals, and domain-specific conferences. View complete publication list on Google Scholar → | Download full CV for details →