Pragnya Sudershan Nalla

Papers

Total Citations

H-Index

About

Pragnya Sudershan Nalla is a leading researcher at the forefront of energy-efficient AI hardware, specializing in composable chiplet architectures and domain-specific accelerators for deep learning. Her seminal work, "CLAIRE: Composable Chiplet Libraries for AI Inference" (2025), introduces a groundbreaking modular framework that enables scalable, high-performance AI inference by assembling pre-verified chiplets, directly addressing the unsustainable computational demands of models like GPT-4 and LLaMAv3. This contribution has already garnered 4 citations, reflecting its immediate relevance in the chip design community. Nalla’s research uniquely bridges the gap between algorithmic innovation and physical hardware constraints, proposing novel methods to reduce power consumption while maintaining throughput for real-time applications in computer vision and NLP. Her work is pivotal in advancing heterogeneous integration, offering a path toward democratizing AI by making powerful inference accessible without monolithic chip costs. With a focus on practical, deployable solutions, Nalla continues to shape the future of AI infrastructure, earning recognition for her ability to translate complex system-level challenges into elegant, composable hardware libraries that promise to redefine next-generation computing.