Software Engineer & Applied AI Researcher

Thilina Dilanka
Jayasuriya

Software EngineerFull-Stack DeveloperApplied AI Researcher

I build reliable, well-architected software and research applied machine learning, from production systems and cloud infrastructure to deep-learning methods for visual material retrieval.

Sri Lanka Software Engineer Consultant @ Docupath Visiting Lecturer · IIT Sri Lanka MSc Advanced Software Engineering (in progress)

View selected work Get in touch

Profile

Who I am

I'm a software engineer and consultant who enjoys the full arc of building products, from planning architecture and standing up cloud infrastructure to the last mile of code review, performance work, and shipping across web and mobile. I care about systems that are maintainable, measurable, and genuinely fast.

Alongside my engineering practice, I'm pursuing an MSc in Advanced Software Engineering, where my research focuses on hierarchical deep-learning systems for visual material retrieval. I move comfortably between production engineering and applied research, and I like bringing ideas from one into the other.

What I do

Focus areas

The core threads that most of my work runs through.

Full-Stack Engineering

End-to-end product development across Python, PHP, and React/TypeScript, from data layer to interface.

Mobile App Development

Designing and building mobile apps for both iOS (Apple) and Android, from concept through to release.

Architecture & Cloud

Designing system architecture and operating on AWS, with containerized and self-hosted infrastructure via Docker and Kubernetes.

Applied AI / ML Research

Deep-learning research on retrieval systems, including CLIP-based models and LoRA fine-tuning for efficient adaptation.

Experience

Where I work

Current Docupath Software Engineer Consultant

Full-Stack Engineering & Architecture

I work as a full-stack engineer and consultant, contributing across the product's lifecycle: planning architecture, managing AWS infrastructure, and leading code reviews to keep the codebase healthy and consistent.

Own architectural planning and technical direction for key parts of the product.
Operate and maintain cloud infrastructure on AWS.
Drive code quality through structured, thorough code review.

                30×
                improvement in the efficiency of a core product module through targeted optimization.
              

Education

Academic background

In Progress MSc, Advanced Software Engineering Informatics Institute of Technology (IIT), affiliated with the University of Westminster

Master of Science

Advanced study across software architecture, engineering practice, and applied machine learning, culminating in a deep-learning research thesis.

Completed BSc (Hons), Artificial Intelligence & Data Science Informatics Institute of Technology (IIT), affiliated with Robert Gordon University

Bachelor of Science (Honours)

Graduated with Second Class Honours (Upper Division), on a foundation spanning machine learning, data science, and software engineering that laid the groundwork for both my engineering practice and my research interests.

MSc Thesis

HiST-CLIP: Hierarchical Retrieval of PBR Materials

A hierarchical deep-learning system for retrieving physically-based rendering (PBR) materials, built on a LoRA-adapted CLIP model to combine strong visual understanding with efficient, parameter-light fine-tuning.

Teaching & Mentorship

Visiting lecturer & supervisor

I serve as a visiting lecturer at the Informatics Institute of Technology (IIT), Sri Lanka, teaching and mentoring undergraduates alongside my engineering work. I also supervise and evaluate BSc final-year students, reviewing their project proposals and guiding their work through to completion.

Many of these projects tackle real, locally-grounded problems in the Sri Lankan context, spanning domains such as tax law, news, and vehicle imports, which makes the mentorship both technical and genuinely applied.

Selected Projects

Things I'm building

A mix of research, product, and personal engineering work.

HiST-CLIP

Research

Hierarchical deep-learning system for retrieving PBR materials, built on a LoRA-adapted CLIP model. My MSc thesis project.

Deep Learning CLIP LoRA Retrieval

Padi.lk

Completed

A completed tech-salary transparency platform for Sri Lanka (also known as Watup.lk), built as a university project and deployed on Kubernetes.

Full-Stack Kubernetes Data

Personal Finance Tooling

Ongoing

An ongoing interest in personal-finance management: designing custom ledger systems and stock-tracking tools to make sense of the numbers.

Ledgers Stock Tracking Tooling

Toolbox

Skills & technologies

Languages

Python PHP TypeScript JavaScript

Frontend

React TypeScript

Mobile

iOS (Apple) Android

Cloud & Infrastructure

AWS Docker Kubernetes Portainer Linux Home Lab Self-hosting

AI / ML

Deep Learning CLIP LoRA Fine-tuning Information Retrieval

Practices

Architecture Design Code Review Performance Optimization

Platforms & Environment

Windows Linux macOS Claude Code

Beyond Work

Away from the keyboard

Outside of engineering I'm an anime enthusiast and enjoy single-player, story-driven games, the kind that lean on world-building and cinematic pacing. I'm also drawn to the craft behind them, and occasionally muse about trying my hand at solo game development.

That same maker's instinct shows up in the small tools I build for myself, from custom ledgers to stock trackers: little systems that make everyday life a bit more legible.