G Abarajithan

PhD student @ UC San Diego working on Hardware Security & Verification, Custom Hardware for Edge AI and SoC-level Integration

CGRA4ML: A Hardware/Software Framework to Implement Neural Networks for Scientific Edge Computing

An open-source, modular framework generating a parameterizable Coarse-Grained Reconfigurable Array (CGRA) in synthesizable SystemVerilog RTL, tailored to common ML compute patterns for scientific applications. G Abarajithan, Z Ma, R Munasinghe, F Restuccia, R Kastner. ACM Transactions on Reconfigurable Technology and Systems, 2026. [Blog Post] [GitHub]

Kraken: An Efficient Engine with a Uniform Dataflow for Deep Neural Networks

A hardware architecture and engine engineered to efficiently process convolutional layers, fully-connected layers, and matrix products within any DNN using a hardware-friendly uniform dataflow designed for maximum data reuse. G Abarajithan, CUS Edussooriya. arXiv preprint. [Blog Series] [GitHub]

Collaborations

Tailor: Altering Skip Connections for Resource-Efficient Inference

A codesign tool that utilizes a hardware-aware training algorithm to remove or shorten skip connections in fully trained neural networks to drastically reduce hardware costs such as BRAMs, FFs, and memory bandwidth. O Weng, G Marcano, V Loncar, A Khodamoradi, G Abarajithan, et al. ACM Transactions on Reconfigurable Technology and Systems 17 (1), 1-23, 2024. [Related Blog Post]

Within-Camera Multilayer Perceptron DVS Denoising

Digital logic implementations of a lightweight multilayer perceptron denoising filter (MLPF) for Event Cameras, successfully reducing the background noise rate by a factor of 100X at the edge. A Rios-Navarro, S Guo, G Abarajithan, K Vijayakumar, et al. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2023. [Blog Post] [GitHub]

Machine Learning on Heterogeneous, Edge, and Quantum Hardware for Particle Physics (ML-HEQUPP)

A community-driven white paper identifying and prioritizing research opportunities in hardware-based ML systems for particle physics, exploring FPGAs, Edge AI, and Quantum Machine Learning. J Gonski, J Ott, S Abbaszadeh, S Addepalli, M Cremonesi, J Dickinson, et al. Whitepaper.

A mostly-online CAS teaching experience

Practical challenges, adjustments, and methodologies in remotely teaching Circuits and Systems (CAS) at undergraduate level in three continents during the COVID-19 pandemic. C Wijenayake, K Wickremasinghe, G Abarajithan, A Madanayake, et al. 2022 IEEE International Symposium on Circuits and Systems (ISCAS), 1783-1787, 2022.