Deniz Najafi · Ph.D. Candidate, NJIT
I design near-sensor and in-memory computing architectures that bring AI inference to the edge, combining integrated photonics, emerging memory technologies, and ViT-class neural accelerators. Advised by Prof. Shaahin Angizi at the ACAD Lab.
I'm a Ph.D. candidate in the Department of Electrical and Computer Engineering at the New Jersey Institute of Technology, working at the ACAD Lab under Prof. Shaahin Angizi. My research builds hardware that closes the gap between AI workloads and the physical devices that run them.
Concretely, I architect near-sensor and in-memory computing accelerators — using silicon-photonic micro-ring resonators, ReRAM/MeFET crossbars, and SRAM/DRAM PIM — to make vision transformers and neuro-symbolic models efficient enough for sensors and IoT nodes. My work spans the full stack: device-level Verilog-A modeling, circuit and array design in Cadence/HSPICE, and architecture-level evaluation for energy, latency, and accuracy.
I'm a Semiconductor Research Corporation (SRC) Research Scholar and a DAC Young Fellow, with first-author papers at ISVLSI and COINS, contributions to DATE, DAC, CVPR, ICCAD, and GLSVLSI, and a recent DATE 2026 Best Paper Award for in-sensor ViT acceleration. I'm currently exploring industry internships in photonic / hardware AI accelerator R&D.
Hardware-software co-design for energy-efficient AI at the edge.
Photonic and electronic accelerators that compute directly at the pixel — eliminating data conversion and movement overhead for vision workloads.
Micro-ring resonators, photonic crossbars, and Verilog-A compact models for end-to-end optical inference pipelines in Lumerical / Ansys.
SRAM, DRAM, ReRAM, and MeFET-based processing-in-memory architectures for CNN and ViT acceleration on TSMC 65nm and FinFET 7nm.
Hardware-aware ViT compression and HyperDimensional Computing for transparent, efficient reasoning on resource-constrained devices.
First-author and lead-collaborator papers. Full list on Google Scholar.
Academic and research milestones along the way.
Application Design Track · April 2026
Best Presentation Award · IEEE/ACM DAC
2024
NJIT ECE · January 2024
College of Engineering GSA Research Day · Fall 2023
First-Year Engineering · 2024 & 2025
