Lucy Revina
Papers
1
Total Citations
5
H-Index
1
About
Lucy Revina is a rising star in computer architecture, whose work bridges the gap between efficient hardware design and the growing demands of machine intelligence. Her research centers on heterogeneous system-on-chip (SoC) architectures, with a particular focus on near-memory computing and domain-specific accelerators for language model inference and robotics. Revina’s most notable contribution is the introduction of NeCTAr (Near-Cache Transformer Accelerator) and RASoC, a pair of 16nm RISC-V class SoCs that integrate both near-core and near-memory accelerators to handle sparse and dense machine learning kernels. Her prototype chip, operating at 400MHz at 0.85V, achieves 109 GOPS for matrix-vector multiplications—a critical operation for transformer-based models. Though early in her career, with her flagship paper already garnering 5 citations, Revina’s work represents a tangible step toward energy-efficient, real-time AI processing at the edge. Her achievements demonstrate a rare ability to translate theoretical acceleration strategies into working silicon, positioning her as a key innovator in the next generation of intelligent, low-power computing systems.
Research Focus
Key Achievements
Top Papers
- 1