KAVINDA.M // BOOT · press any key to skip

SIGNAL CHAIN · STAGE 1 / SILICON

Embedded Systems Architect ·

I engineer the full stack of computing - from transistors to the cloud.

Signal out→
Scroll

Transistors alone do nothing — wire them into gates and they compute.

◢ Signal Chain · Stage 2 / Logic
0+
0000000
Years · Silicon → Cloud
0
0000000
px LED Rig · One Controller
0
0000000
ArtNet Universes @ 30 FPS

Engineering Electronics & Experiences

Since 2019

I work the full computing stack — from transistor-level logic and custom assembly, through firmware and embedded Linux, up to networking and cloud. That range was built on years of motherboard-level diagnostics and reverse engineering, giving me a practical feel for systems from silicon to software.

What sets me apart is bridging deep hardware expertise with real software architecture: designing deterministic, high-throughput systems for live multimedia — DMX and laser control, NDI streaming, real-time LED pipelines — as well as scalable multi-tenant SaaS.

Currently reading an MSc in Computer Science (Software Architecture) at the University of Moratuwa, connecting low-level systems engineering with scalable enterprise design.

Published AI research · motherboard fault diagnosis
Oman Muscat Festival 2025 · AV & IP infra
Built an LED engine: one RP2350 driving 12,288 pixels over 72 ArtNet universes
Toolkit · Silicon → CloudSTACK
Languages
CC++Assembly (x86 · RISC-V)PythonTypeScriptBash
Hardware & MCUs
STM32ESP32RP2040 / RP2350RK3588JTAGPCB Design
OS & Kernel
Embedded LinuxFreeRTOSOpenWrtDevice DriversDevice Tree
Multimedia & Networking
GStreamerNDIArtNet / DMXCAN BusMQTTWebSockets
Architecture & Cloud
DockerPostgreSQLNext.jsNestJSCI/CD
ALU · Logic Unit4-BIT · GATE LEVEL
A
11
B
6
=
1
signed 1

click a result bit to inspect its gate slice →

CARRY
ZERO
NEG
OVF
BIT 0 · FULL ADDER GATE SLICE
A01B00Cin0XORANDXORANDOR10SumCout
DATAPATH · 4× FULL ADDER · RIPPLE CARRY — click a slice to inspect it
CoutA31B30FAbit 3c=10S3=0A20B21FAbit 2c=10S2=0A11B11FAbit 1c=00S1=0A01B00FAbit 0Cin=01S0=1
Same operation, up the stack
Assembly · what the CPU runs
MOV R0, #11 ; 0b1011
MOV R1, #6 ; 0b0110
ADDS R2, R0, R1 ; R2 ← R0 + R1
R2 = 0b0001 (1)flags: N Z C V
C · what the engineer writes
uint8_t a = 11; // 0b1011
uint8_t b = 6; // 0b0110
uint8_t c = a + b; // c = 1
same gates, same carries — the compiler just hides them

Once hardware can compute, someone has to tell it what to do.

◢ Signal Chain · Stage 3 / Firmware
/dev/ttyUSB0 · 115200 8N1idle
$ make flash-toolchain
Services · What I Do4 modules loaded

Services I Offer

Comprehensive solutions tailored to your needs, delivered with expertise and passion.

Software Engineering

End-to-end software solutions, from embedded systems to full-stack applications. Specializing in microcontroller programming, FPGA development, and multimedia systems.

View Details

Hardware Repairs

Specialized in Apple device motherboard repairs and high-end chip-level diagnostics. Expert in iPhone and MacBook logic board fault diagnosis and repair.

View Details

Tech Education

Personalized and group training in ICT, electronics, and programming. Hands-on experience with FPGA, IoT, and practical engineering projects.

View Details

Technical Consulting

Helping businesses and individuals with technology planning, troubleshooting, and strategy. From architecture design to system optimization.

View Details

Software Engineering

Premium Service

End-to-end software solutions, from embedded systems to full-stack applications. Specializing in microcontroller programming, FPGA development, and multimedia systems.

What's Included

▸Embedded C/C++ & Python Applications
▸IoT System Design
▸FPGA Programming
▸API & Integration Development
▸Multimedia Systems
▸Custom Automation Solutions

This is what my firmware does for a living: turn network packets into light.

◢ Signal Chain · Stage 4 / Signal

The Signal Chain — my specialty, end to end

Live-event lighting and video is a hard-real-time problem: protocol in, processing in the middle, photons out — every 33 ms, without fail. I've built every link of this chain in production.

01 · Input · Protocols

Control and video arrive on the wire — lighting consoles, media servers, capture.

  • ArtNet / DMX / sACN

    72 universes patched — zero frame loss @ 30 FPS

02 · Processing · Pipelines

Media engines and custom plugins shape every frame in real time.

  • GStreamer / NDI / HDMI Capture

    Media pipeline negotiated — NDI stream live

  • TouchDesigner C++ Plugins

    KineticCHOP loaded — 80ch DMX motor control

03 · Output · Light

Deterministic firmware turns frames into photons — LED walls and laser galvos.

  • LED Systems (WS2812 @ scale)

    12288 pixels mapped — DMA ring buffers primed

  • Laser ILDA Control

    Galvo calibrated — ILDA trace armed

Simulation · The real rig I built

RP2350 dual-core · 72 ArtNet universes · 12,288 WS2812 pixels at 30 FPS. This canvas recreates it pixel-for-pixel — click to inject a pulse.

UNIVERSES72
PIXELS12,288
REFRESH30 FPS
Featured Channels · Deep Dives

One signal chain is a demo — here are the full products it lives inside.

◢ Signal Chain · Stage 5 / Systems
Project Index · 6 systemsEmbedded → Web

Featured Projects

Complete systems built up the stack — firmware, media platforms, and full-stack products.

Featured
Elites Institute LMS

Elites Institute LMS

A comprehensive Learning Management System designed for Grade 6 to A/L students in Sri Lanka. The platform enables seamless enrollment for ICT and Music classes, offering features like student registration, class scheduling, and resource distribution. Built to modernize traditional education with digital accessibility.

Next.js
React
Tailwind CSS
Django
PostgreSQL
+1 more
FrameLight Studio

FrameLight Studio

A professional portfolio site for a photography and videography studio. Showcasing expertise in capturing weddings, birthdays, graduations, and special events. The platform highlights their 'frame by frame' storytelling approach and service offerings.

React
Netlify
iPhone Motherboard Fault Diagnosis Tool

iPhone Motherboard Fault Diagnosis Tool

A deep learning-based diagnostic system designed to detect iPhone motherboard faults through power consumption variation analysis. The tool identifies defective components and fault patterns at chip level, enabling faster and more accurate repair decisions for technicians.

Python
TensorFlow
OpenCV
NumPy
Matplotlib
+1 more
Rivisara Drum Institute

Rivisara Drum Institute

A digital presence for Sri Lanka's premier traditional drum institute. The platform facilitates student enrollments for drumming and dancing courses, showcases a gallery of cultural performances, and allows clients to book services for weddings and corporate events. It serves as a bridge between traditional heritage and modern convenience.

Next.js
React
Tailwind CSS
Supabase
Vercel
StudioFlow SaaS

StudioFlow SaaS

A premium SaaS platform specially engineered for wedding photographers to streamline their business. Features a dedicated Client Portal for tracking progress, a masonry Gallery for showcasing portfolios, and a robust Admin Dashboard for managing projects and deliveries. Designed to professionalize the photography workflow.

Next.js
TypeScript
Tailwind CSS
Clerk
Framer Motion
+2 more
StudioFlow Marketing & Demo

StudioFlow Marketing & Demo

The public-facing marketing website and interactive demo for the StudioFlow SaaS platform. It showcases the platform's capabilities to potential users, featuring high-performance animations, optimized imagery, and a compelling user journey to convert visitors into subscribers.

Next.js
React
Tailwind CSS
Framer Motion
PostgreSQL
+2 more

Every device I ship reports to something — this is that something.

◢ Signal Chain · Stage 6 / Cloud
Cloud Systems · Deep Dive

Research & Publications

Development of an Intelligent Fault Diagnosis Tool for iPhone Motherboards: Power Consumption Analysis Using Deep Learning

P.D.K. Madhubhashana, H.D.N.V. Jayasekara, G.D.G.N. Jayawardena, B.N.S. Lankasena, B.M. Seneviratne

Advances in Artificial Intelligence and Machine Learning (OAJAI&ML), Vol. 5(2), pp. 3784-3808 • 2025

DOI: 10.54364/AAIML.2025.52215

Abstract

This study presents an intelligent microcontroller-based diagnostic tool and application designed to enhance fault detection accuracy and efficiency in iPhone motherboards, utilizing power consumption data and deep learning for real-time diagnostics. The tool, deployed in phone repair centers, has generated a comprehensive dataset of over 1,600 iPhone 6s devices with faults linked to 12 distinct power rails.

Key Highlights

  • Power profile analysis during boot and operational states
  • LSTM-based fault classification model trained on real-world data (99% accuracy)
  • Diagnostic accuracy surpassing traditional fault detection methods
  • Dataset of 1,600+ iPhone 6s devices across 12 power rails
  • RP2040 microcontroller with INA226 current sensor integration
  • Potential applications in advanced mobile repair labs and refurbishing industries
View Full Publication

End of the chain. Your move.

◢ Signal Chain · Stage 7 / Contact

Let's Connect

Have a project in mind? Let's discuss how we can work together.

ILDA vector trace · galvo simulation · signal out

Loading contact form...