Integrating Multiscale Modeling and Differentiable Physics for Computing and Energy Applications
My research bridges first-principles calculations, differentiable models, mesoscale simulations, and manufacturing-aware design to enable high-performance solutions in computing and energy systems.

Recent news

05.21.25 Giuseppe gave an invited talk the International Workshop on Computational Nanotechnology (IWCN), Salt Lake City UT (USA) entitled “Differentiable Phonon Hydrodynamic Simulations for Inverse Design and Material Property Extraction.”

04.17.25 Giuseppe gave an invited lecture (virtual) in the Simulation Class at the Georgia Institute of Technology entitled “Accelerated Materials Discovery with Differentiable Programming: From Thermal Energy Harvesters to Chiplet Design.”

04.15.25. Giuseppe gave a colloquium at the Department of Electrical Engineering of Boston University, entitled “Accelerated Materials Discovery with Differentiable Programming: From Thermal Energy Harvesters to Chiplet Design” (abstract)

03.31.25. Our paper “Analog-Computing with Heat: Matrix-vector multiplication with inverse-designed metastructures” has been posted on arXiv.

03.11.25. Our conference paper for the 2025 Electronic Components and Technology Conference (ECTC 25) entitled “DiffChip: Thermally Aware Chip Placement with Automatic Differentiation,” has been posted on arXiv. Great collaboration with the MIT-IBM Watson AI Lab!

02.27.25. We received the 2024 Raith Award for the best (second place) nanostructure patterned with the Velion Focused Ion Beam (FIB). The structure is an inverse-designed nanostructure for thermal energy harvesting (thermoelectrics), with a feature size of about 100 nm.