Federal Project Grant Award Summary

Carnegie Mellon University received a $395,936 CAREER award from the National Science Foundation (NSF) Division of Information and Intelligent Systems under the Computer and Information Science and Engineering (CISE) program (CFDA 47.070) for the period May 1, 2026, through April 30, 2031. The award supports the development of an automated machine learning (ML) compiler toolchain designed to optimize and accelerate artificial intelligence model deployment across diverse computing environments, including data centers, mobile devices, and web browsers. The project will deliver a unified program representation that encapsulates high-level computational encodings and low-level optimizations, coupled with an artificial intelligence-driven ML compiler agent that leverages structured tools to iteratively optimize deployment performance.

The core products under this award include: (1) a unified methodology for representing optimizations across the ML stack, spanning from high-level modeling choices to low-level hardware-specific optimizations; (2) an AI-based compiler system that integrates domain expertise to automate optimization decisions; and (3) end-to-end solutions enabling rapid deployment of advanced ML models to emerging platforms and hardware architectures. The toolchain will enhance accessibility of ML models across cloud and edge computing environments, improve the broader ML tools ecosystem supporting commercial applications, and reduce engineering effort required to establish ML deployment infrastructure for new hardware technologies. No subawards are planned for this project.