I build enterprise systems that hold up under real load — and I use AI to get them there faster.
Software Engineer at Innovative-e, Inc., working on large-scale distributed systems for a global client base. BSc (Hons) Computing, Coventry University.
Short version
I'm a Software Engineer at Innovative-e, a Microsoft partner in project and portfolio management, where I work on large-scale distributed systems for a global client base. My job is mostly translation: taking a business requirement that arrives vague and expensive, and turning it into architecture that is specific, testable and cheap to run.
The other half of my work is AI. Not as a demo, but as infrastructure — integrating models into the SDLC to shorten delivery cycles, automating the workflows that quietly eat a team's week, and building agentic systems with LangChain, n8n and CrewAI. I run local models too, because some problems shouldn't leave the building.
Where I've built things
Software Engineer
Innovative-e, Inc. — Colombo, Sri Lanka
Leading multiple high-impact enterprise web projects with a focus on scalability, security and high-availability architecture.
- Architect and deliver end-to-end full-stack solutions for international enterprise clients
- Integrate AI into the delivery lifecycle to compress cycle time and automate manual workflows
- Own projects from requirements analysis through to cloud deployment and operation
Associate Software Engineer
Innovative-e, Inc. — Colombo, Western Province, Sri Lanka
Built and shipped features across the Microsoft platform stack for global project-and-portfolio-management clients.
- Delivered production features across .NET, React and the Power Platform
- Worked directly with international stakeholders on requirements and rollout
Software Engineering Trainee
Innovative-e, Inc. — Colombo District, Western Province, Sri Lanka
Moved from supervised tasks to owning discrete features in client-facing systems.
Software Engineer — Intern
Innovative-e, Inc. — Colombo, Western Province, Sri Lanka
First professional engineering role. Learned the enterprise delivery process end to end.
Full-stack Developer
TIOSS — Remote
Built full-stack web features for client projects on a distributed team.
Freelance Web Developer
TIOSS — Sri Lanka
Delivered web projects for local businesses — design through deployment.
Marketing Team Member
AIESEC — NIBM Colombo, Sri Lanka
Ran campaigns and content for the local chapter alongside full-time study.
Education
- Jan 2021 — Dec 2025
BSc (Hons) Computing
Coventry University
Computer Software Engineering
- Jan 2021 — Dec 2022
Advance Certificate in Software Quality Assurance
National Institute of Business Management (NIBM), Sri Lanka
Computer Software Engineering
- Jan 2019 — Dec 2020
Diploma in Information Technology
ESOFT Metro Campus
Information Technology
What I work in
Listed honestly: these are things I have shipped with, not things I have read about. The AI column is the half of my practice that has changed most in the last two years.
Engineering
06- C#
- Java
- Python
- C / C++
- PHP (Laravel)
- TypeScript
Web
04- React
- Next.js
- Tailwind CSS
- WordPress
Data
05- MS SQL Server
- PostgreSQL
- MongoDB
- Cosmos DB
- Supabase
DevOps & QA
05- Azure DevOps
- CI/CD
- Selenium
- TestNG
- Postman
AI Development
06- Claude Code
- Cursor
- Devin
- Antigravity
- Lovable
- Z AI
Agentic & Local
06- LangChain
- CrewAI
- n8n
- OpenClaw
- Ollama
- Jan
Languages
- EnglishProfessional working proficiency
- SinhalaNative or bilingual proficiency
Certifications
- C++
- AI-Powered Web Development
Apps
TokenLens
Token, character and word counts as you type — in every AI chat.
A browser extension that counts tokens live inside Claude, ChatGPT, Gemini, Grok and Kimi. Everything runs locally: no API keys, no server, no network requests. GPT counts are exact via the public o200k_base tokenizer; the rest are estimated with per-script calibration, which matters more than most tools admit — Sinhala costs roughly three and a half times more tokens per character than English, and a flat "4 characters per token" rule is badly wrong for it.
- Exact GPT token counts with a locally bundled tokenizer
- Per-script calibration for Latin, Sinhala, CJK and code
- Live in-page counter on five major AI chat sites
- Fully offline — no keys, no telemetry, no requests
Writing
I'm interested in hard problems at scale — distributed systems, AI integration that actually ships, and teams who care about the second year of a codebase. If that's you, say hello.