Alfio Di Mauro
Electronic engineer, SoC researcher and startup founder
Zurich, Switzerland
Summary
Expert in low-power SoC and edge-AI hardware: Alfio's work centers on designing energy-efficient system-on-chip architectures and accelerators for event-driven and neuromorphic computing. His publications and project contributions (SNE, Kraken and several IEEE/journal papers) demonstrate deep expertise in combining RISC-V clusters, mixed-precision DNN accelerators, and neuromorphic engines for ultra-low-power vision and AI tasks. arxiv+2
Academic researcher and PhD-trained systems designer: Alfio completed doctoral research at ETH Zurich (2017–2022) on system-on-chip architectures for event-driven computing and has an extensive publication record and thesis documenting his contributions to energy-efficient embedded processing. ethz+1
Active contributor within the PULP open-source ecosystem: He has been part of the PULP Platform effort since 2017, collaborating on multi-author SoC projects and software stacks that target constrained embedded devices and nano-UAV applications. pulp-platform+1
Entrepreneurial founder moving research to product: Alfio co-founded Mosaic SoC and serves as CEO, aiming to commercialize low-power perception SoCs and software for AR/VR, wearables and robotics while engaging with investors and industry partners. venturelab+2
Work
Education
Projects
Writing
SNE: an Energy-Proportional Digital Accelerator for Sparse Event-Based Convolutions
April 1, 2022Paper presenting a 4.5 TOP/s/W digital accelerator for 4-bit quantized event-based CNNs that performs operations proportional to input event activity, achieving very low energy per operation.
Kraken: A Direct Event/Frame-Based Multi-sensor Fusion SoC for Ultra-Efficient Visual Processing in Nano-UAVs
January 1, 2022Paper describing Kraken, a heterogeneous ultra-low-power SoC fabricated in 22nm FDX that fuses event-based and frame-based visual sensors and integrates multiple accelerators for on-board perception on nano-UAVs.
Doctoral thesis — System-on-chip Architectures for Event-Driven Computing
January 1, 2022PhD thesis covering architectures and system-level design for event-driven computing and low-power SoC implementations, accompanied by multiple conference and journal publications.