Two Indian American researchers at Duke University, Gaurav Arya and Tania Roy, have been selected to lead components of four Duke-affiliated teams chosen for the U.S. Department of Energy's Genesis Mission, a sweeping national initiative aimed at using artificial intelligence to accelerate scientific discovery across a range of fields — from modelling atomic nuclei and reacting faster to stellar explosions, to imbuing robots with more brain-like circuitry and designing entirely new DNA-based materials. The Genesis Mission is being built by the DOE as what the agency describes as the world's most powerful integrated science discovery platform, and the newly announced Phase I awards, ranging from $500,000 to $750,000 and running for nine months, mark an early but significant step in that ambitious federal undertaking.
Arya, a professor in Duke's Pratt School of Engineering within the Thomas Lord Department of Mechanical Engineering and Materials Science, leads the Duke project on 'AI-Driven Inverse Design of Patchy DNA Origami for Assembly of Programmable Superlattices.' The project centres on an emerging and genuinely cutting-edge area of materials science: DNA origami, a technique that involves precisely folding DNA strands to create materials with nanoscale structures engineered for specific energy-related applications. While folding DNA in different configurations can, in principle, produce an enormous variety of distinct materials, actually planning and designing those configurations represents an extraordinarily difficult computational challenge — one that Arya's team believes artificial intelligence is particularly well positioned to help solve.
According to Arya, the research effort is aimed squarely at an enormous, largely untapped design space, with the team expecting that the novel biomaterials emerging from this work could meaningfully impact industries ranging from energy production and chemical manufacturing to, potentially, quantum computing. That breadth of potential application reflects just how foundational this kind of materials-design research can prove to be — DNA origami-derived materials, if the AI-driven design methodology Arya's team is developing proves successful, could feed into multiple, otherwise unrelated downstream industries simultaneously, a characteristic that likely factored into the Genesis Mission's decision to fund the project as part of its broader scientific discovery platform.

Arya's academic research more broadly focuses on using physics-based computational tools to build fundamental, molecular-level understanding of a diverse range of biological and soft-material systems, with the explicit aim of discovering new phenomena and developing new technologies from that foundational understanding. That research orientation — grounding highly applied, technology-oriented outcomes in rigorous, first-principles physical modelling — reflects a research philosophy that has become increasingly influential across materials science and computational biology over the past decade, as advances in computing power have made physics-based simulation an increasingly viable complement to, and sometimes substitute for, purely experimental materials discovery.
Tania Roy, an Associate Professor of Electrical and Computer Engineering also within Duke's Pratt School of Engineering, is part of a second Duke project, led by fellow Duke researcher Yiran Chen, focused on 'Neuromorphic Circuit Primitives for Robotic Embodied Physical AI.' The project's central goal is to develop new AI hardware for robots that the team says could prove up to ten times faster and 100 times more energy-efficient than current robotic AI hardware designs — a dramatic potential leap in both speed and efficiency that, if achieved, could meaningfully expand the practical range of tasks AI-powered robots are capable of performing outside of narrow, highly controlled niche applications.
The technical approach underpinning Roy and Chen's project involves designing neuromorphic processors — computing hardware deliberately structured to mimic the architecture and function of biological brains and nervous systems, rather than following the more conventional digital computing architectures that have dominated electronics for decades. That neuromorphic approach directly targets what the research team has identified as one of the central bottlenecks currently limiting AI-powered robotics: the speed at which robotic systems can process incoming data from cameras and other sensors, a processing bottleneck that has, according to the team, kept many otherwise promising AI robotics applications confined to narrow, specialised use cases rather than broader, more general-purpose deployment.
Roy's own research programme focuses specifically on developing hardware for artificial intelligence applications built around novel functional materials, including two-dimensional materials — an active and rapidly evolving subfield of materials science and electrical engineering that has drawn increasing research investment globally as the limitations of conventional silicon-based computing architectures have become more apparent, particularly for the kind of power-hungry, computationally intensive AI applications that have proliferated over recent years. Chen, the lead researcher on the neuromorphic robotics project, offered a particularly vivid description of the team's ultimate technical ambition, describing the eventual completed system as resembling a synthetic organism whose fundamental neuromorphic computing components would function much like muscles, skeleton, and nervous system working together in a biological body.




