kanishka@portfolio ~
$ |

Hello, I'm

Kanishka Gunawardana

Computer Engineering Graduate

4.0 GPA
#1 Rank / 486
3 Publications
10+ Projects
Accelerated Computing Computer Architecture Embedded Systems Neuromorphic Computing

I specialize in computer architecture and embedded systems, designing hardware–software co-designs that optimize CPU microarchitecture and maximize performance. My work advances energy efficiency and computational throughput in real-world embedded and edge applications.

Focus Areas

Exploring the intersection of computer architecture, neuromorphic computing, and intelligent systems for next-generation computing platforms.

Computer Architecture

Hardware-Software Co-Design Domain-Specific Accelerators RISC-V SoC

Designing and verifying high-performance processor architectures and domain-specific accelerators. I leverage hardware-software co-design to optimize performance, including implementing RISC-V (RV32IM) pipelines and integrating them into complex SoC designs.

Embedded Systems

Low-Power Edge AI IoT Real-time Systems

Creating robust and intelligent embedded solutions for real-world applications. My expertise covers the design and implementation of real-time systems and low-power hardware, including firmware development and centralized hub design.

Neuromorphic Computing

Brain-inspired Computing Spiking Neural Networks Edge AI

Developing brain-inspired computing systems for next-generation Edge AI. My work with the PeraMorphIQ research group focuses on designing configurable neuromorphic accelerators, optimizing for state-of-the-art energy efficiency.

Education

Academic journey with exceptional performance in computer engineering and mathematics.

University of Peradeniya

Nov. 2021 – Present

Computer Engineering B.Sc. Engineering (Hons.) First Class

Undergraduate in B.Sc. Engineering (Hons.) Computer Engineering

GPA: 4.0/4.0
Rank: 1st out of 486 students

Dharmaraja College Kandy

Nov. 2006 – Aug. 2019

G.C.E. Advanced Level Mathematics Stream Top 150 Nationally

G.C.E. Advanced Level Examination with outstanding performance

Z-score: 2.5661
National Rank: 149 out of 19,508
District Rank: 11 out of 1,189

Professional Experience

Academic and industry experience in computer engineering, research, and software development.

Research Work

Ongoing and completed research projects in computer architecture and neuromorphic computing.

SNAP-V: A RISC-V SoC with Configurable Neuromorphic Acceleration for Small-Scale Spiking Neural Networks

Nov. 2024 - Jul. 2025 | Final Year Thesis

Final Year Thesis RISC-V SoC Neuromorphic H-NoC Energy-Efficient

Designed and developed a dual-core RISC-V System-on-Chip integrating a configurable neuromorphic accelerator with over 1k LIF neurons organized into parallel clusters interconnected through a hierarchical Network-on-Chip (H-NoC).

Validated the SoC on MNIST using Synopsys (VCS/PrimePower) and Xilinx Vivado, achieving 96.69% accuracy and state-of-the-art energy efficiency of 1.39 pJ/synaptic operation.

Supervision: Dr. Isuru Nawinne, Prof. Roshan G. Ragel

Featured Projects

Highlighting significant contributions to research and development in computer engineering.

Ballerina OpenAI Finetunes Connector

Aug. 2024 - Sep. 2024

Open-Source WSO2 Internship Ballerina OpenAI API

Developed the ballerinax/openai.finetunes connector during WSO2 internship, providing seamless access to OpenAI's fine-tuning API through Ballerina.

Impact: Published on Ballerina Central with full documentation.

Publications

Peer-reviewed publications and contributions to academic conferences.

SNAP-V: A RISC-V SoC with Configurable Neuromorphic Acceleration for Small-Scale Spiking Neural Networks

Preprint RISC-V SoC Neuromorphic Edge AI

Authors: K. Gunawardana, S. Peeris, K. Rambukwella, T. Wanduragala, S. Jameel, R. Ragel, I. Nawinne
Status: Preprint (arXiv)

  • Proposed SNAP-V, a RISC-V-based neuromorphic SoC featuring two accelerator variants (bus-based Cerebra-S and NoC-based Cerebra-H) to provide a complete, resource-efficient platform tailored for small-scale edge SNNs.
  • Demonstrated a 9.46x operating frequency improvement in Cerebra-H over Cerebra-S, achieving 1.05 pJ/SOP in 45nm CMOS while maintaining high inference fidelity with a 2.62% average accuracy deviation.

Neuromorphic Architectures for Edge-Oriented Spiking Neural Networks: A Review

Preprint Review Paper Neuromorphic Computing SNNs

Authors: K. Gunawardana, S. Peeris, K. Rambukwella, R.G. Ragel, I. Nawinne
Status: Preprint (SSRN)

  • Conducted a review of ASIC- and FPGA-based neuromorphic architectures, highlighting large-scale systems are underutilized for small-scale edge workloads while small-scale designs remain rare; analyzed software frameworks and event-driven communication strategies to advocate for SoC-centric platforms for energy-efficient deployment.

Optimized Multi-Processor System-on-Chip (MPSoC) Design for Low-Resource JPEG Encoding

IEEE ICAC 2024 MPSoC FPGA Custom Instructions

Authors: K. Gunawardana, C. Adhikari, I. Nawinne
Conference: IEEE International Conference on Advanced Computing (ICAC) 2024, Colombo

  • Implemented an FPGA MPSoC JPEG encoder on Cyclone IV using Nios II/e cores; introduced lightweight custom hardware instructions and custom FIFO queues to offload compute-critical stages and reduce processor stalls.
  • Achieved 2.8× throughput improvement; superscalar options were evaluated but deprioritised.

Recognition & Achievements

Awards and recognitions that reflect excellence in academics, research, and competitive programming.

J. B. Dissanayake Prize

2025 | Faculty of Engineering, UoP

Industrial Training Excellence
Awarded for outstanding performance and excellence during the undergraduate industrial training program.

SLIot Challenge 2023

Mar. 2024 | Team: IMPAX

1st Runners-up
Sri Lanka's biggest IoT competition
(100+ teams)

IEEEXtreme 17.0

Nov. 2023 | Team: Five4Five

Global Rank 374
24-hour programming competition
(16,500+ participants)

MoraXtreme 8.0

Nov. 2023 | Team: Five4Five

4th Place National
12-hour algorithmic programming
(400+ teams)

ACES Coders v10.0

Oct. 2023 | Team: Five4Five

12th Place National
12-hour algorithmic programming
(350+ participants)

ACES PreCoders v10.0

Sep. 2023 | Team: Five4Five

2nd Place University
6-hour algorithmic programming
(50+ teams)

ACES Hackathon 2023

Sep. 2023 | Team: LearnLink

Participant
Inter-university hackathon
LearnLink - Online marketplace for books

Technical Skills

Comprehensive technical expertise across hardware design, software development, and modern engineering tools.

Programming Languages

Python C/C++ Verilog HDL ARM Assembly Java JavaScript TypeScript SQL Ballerina PHP

Frameworks & Platforms

Arduino ESP32 Express.js Spring Boot FastAPI Node.js React.js

Libraries & Tools

OpenCV NumPy Pandas Matplotlib PyTorch TensorFlow

Developer Tools

Git Linux Docker AWS

EDA & Verification Tools

Quartus II Nios II GTKWave Vivado Synopsys Design Compiler VCS PrimeTime PrimePower

Selected Certificates

Professional development through specialized courses and certifications.

Get In Touch

Interested in collaboration, research opportunities, PhD positions, or engineering roles? Let's connect!

Location

Department of Computer Engineering
University of Peradeniya, Sri Lanka