Samitha Nanayakkara, Senior Technical Lead
senior technical lead · ai agents, web and mobile · colombo
here’s my recent work.
nine agents, four codebases, one app in two weeks.
01 / 03 · 2025 · multi-agent build · 10k+ daily users
a government paper trail, compressed to minutes.
02 / 03 · 2025 · government university · built in 5 days
one ai agent grading another. client satisfaction up 32%.
03 / 03 · 2025 · ai evaluation · +32% satisfaction
and here’s what’s already shipped.
downloads of the Celcom Life super app
2019
saved for Dialog by shipping a feature their last vendor couldn't, without rewriting the app
2019 · Axiata Digital Labs
rise in client satisfaction from an AI eval pipeline
2025
faster PR merges after rolling out AI code review
2025
also: $60K+ yearly cloud savings from on-device encoding · 4.8★ rating on a Mobile App of the Year · 20+ products shipped since 2018
and this is how I make decisions.
- One owner for every part.seen in: nine agents →
- Trust what users do over what dashboards say.seen in: eval pipeline →
- Decide when to stop before you start.seen in: eval pipeline →
- Model the process people already follow.seen in: exam vouchers →
this is how I think about what I build.
- What multi-agent systems taught me that isolated agents didn’tA lesson in collaboration design I learned the hard way
- Leverage > Lines of CodeWhy the engineers I learn most from write the least code
- Your degree isn’t obsolete. Your mindset might beWhat a CS degree still gives you
- Beyond the hype: real gains with AI code reviewAbout 75% faster merges, and how we measured it
- Solving a sneaky Next.js memory leakA debugging story that started with a slow dashboard