Federal Project Grant Award Summary

The University of Texas at Austin received a $300,000 Project Grant from the National Science Foundation's Office of Advanced Cyberinfrastructure under the Computer and Information Science and Engineering program (CFDA 47.070), awarded October 1, 2026, with completion targeted for September 30, 2029. The grant funds development of an AI-based conversational assistant framework designed to simplify the configuration and deployment of scientific computing environments on shared research cloud infrastructure platforms such as Chameleon and Jetstream2. The system enables researchers to describe their computational requirements in natural language rather than requiring specialized expertise in system administration, networking, and security configuration.

The deliverables include an integrated AI agent framework combining large language models with software tools that automatically plan, provision, and validate scientific computing environments. Key components encompass planning modules with resource limit and hardware compatibility checks, state and error handling modules that track multi-step workflows, search pipelines that aggregate information from diverse documentation sources, and validation benchmark suites. This solution addresses the challenge of managing complex scientific workloads involving tightly coupled virtual clusters, specialized hardware, and accelerators, with the intended benefits of increased scientific productivity, reduced cloud computing costs, and improved computational reproducibility for research teams.