Fig. 01 — Senior engineer · Tech lead · Colombo, LK
Senior Laravel engineer & tech lead for backend systems that can’t afford to fail.
I design and build the systems behind the login screen — fintech platforms, multi-tenant SaaS, ERPs and admin panels — for companies that need them correct, fast, and still easy to change two years from now.
Open to new projects · Hourly or fixed price · Remote, worldwide
1
2
3
| Engineer | H. P. Bandara | Discipline | Backend · Architecture · Leadership |
|---|---|---|---|
| Stack | Laravel · PHP · MySQL · PostgreSQL · Redis · Vue.js · AWS | ||
| Since | 2019 | Rev | 2026.11 · Sheet 01/09 |
I’m Harshadeva, a senior full-stack engineer from Colombo, Sri Lanka. I’ve spent seven years building Laravel systems that run in production every day.
Most of my work lives behind a login screen: fintech platforms, ERPs, admin panels, multi-tenant SaaS and the APIs that hold them together. It doesn’t make for flashy screenshots. It does make for systems that businesses depend on, where a wrong number or a double charge is a real problem.
I lead teams, review code and make architecture calls, and I still write code every day. Good technical decisions come from staying close to the work.
What teammates say about working with me
Off the clock
Driving, reading, working out. Long drives clear my head, books keep me curious, and training keeps me disciplined. These are the same three things I bring to engineering: a clear head, curiosity, and the discipline to do it properly.
Most failed projects don’t fail in the code. They fail because people imagined different things and found out too late. My process exists to close that gap before the first line of code, and to keep it closed while we build.
Who uses this? What happens to the business if it breaks? What does “done” actually look like? I ask the uncomfortable questions on day one, so they don’t come up in week six.
Output → problem statement, user roles, constraints, risks
Flow diagrams, data models, sequence diagrams and a written scope. We go back and forth on paper until you, me and the team describe the project the same way. Changing a diagram takes minutes. Changing shipped code takes weeks.
Output → diagrams + scope document everyone has agreed to
I break the work into small pieces that can each be delivered on their own, and estimate each one. Small pieces give honest estimates, and you can see exactly where the time goes.
Output → prioritized backlog with per-item estimates
Agile, in short cycles. At the end of each phase you review working software, not a status report. If what’s being built starts drifting from what you expected, we catch it while it’s still cheap to fix.
Output → a working, reviewable build every phase
Code review, automated checks, CI/CD pipelines and static analysis on every change. Deployments should be boring, and I set them up that way on purpose.
Output → reviewed code, green pipeline, repeatable deploys
Logs, metrics and alerts from day one, so problems are spotted by monitoring and not by your customers. Then we improve based on real usage, not guesses.
Output → dashboards, alerts, a list of improvements backed by data
SOLID
Each part does one job. New features extend the system instead of forcing a rewrite, so the tenth feature costs about the same as the first.
ACID
Money and records move inside transactions: all or nothing, never half. No orphaned orders, no balances that don’t add up.
DRY
Each business rule lives in one place. When it changes, it changes once, and the bug can’t hide in a forgotten copy.
QUALITY > RUSH
When the deadline is tight, I cut scope, not quality. Shortcuts aren’t free. You pay them back later, with interest.
The 2 AM rule
Looks cheap now. It won’t be.
Backend work is hard to show in screenshots, so below are small working models of the problems I solve. Try them.
5.1
I built and scaled a fintech platform that handles around 100 transactions per second across 100+ payment gateways and services, including Stripe.
What matters here are the unglamorous details: idempotency keys, safe retries, duplicate-charge prevention, and webhook handling for success, failure and refunds that keeps every system consistent. These details decide whether a customer gets charged twice.
5.2
I build SaaS platforms where every customer gets a dedicated database, inside a modular monolith.
Each client’s data is physically separated. Each module has clear boundaries. The whole thing still deploys as one application, so you don’t pay the cost of microservices before you actually need them.
GET acme.app.com/invoices
Tenant resolver → connection tenant_acme
5.3
A critical report took 2 minutes to load. After profiling, rewriting the slow queries and adding the right indexes, it loads in 2 seconds.
On the same platform I cut AWS costs by 40%. Slow systems are usually a few bad decisions repeated many times. Find them, fix them, and keep them from coming back.
5.4
Role-based access, multi-level permissions and real-time dashboards built with server-sent events and WebSockets. One ERP I architected serves 4,000+ organizations nationwide.
5.5
Clean REST APIs and dependable third-party integrations: payment gateways, Stripe, Google Maps, Firebase push notifications, SMS and email, and webhooks.
5.6
AWS deployments, Jenkins CI/CD, SonarQube quality checks, and Elasticsearch with Kibana, so you can see what the system is doing.
When two people on a team disagree, I don’t pick a side over chat. I get both of them in the same room, at the same time.
Most of the time, the disagreement goes away once the scope is clear. When it doesn’t, it’s a real trade-off, and we decide it with data, not volume.
Before: two people, two different projects in their heads.
Reviews are how standards spread through a team. I explain the why, not just the what, so the same comment never needs writing twice.
I give junior developers real ownership with a safety net: clear tasks, pairing when they’re stuck, and honest feedback. A team that grows needs less managing.
I turn business requirements into technical plans, and technical risks into business language, so stakeholders and engineers can make decisions together.
Sep 2022 — Present
Webmotech Pvt Ltd · Sri Lanka
Scalable Laravel backends, AWS deployments, CI/CD, code reviews, mentoring and architecture decisions.
Sep 2020 — Sep 2022
Sevensings Digital Solutions · Sri Lanka
Full-stack Laravel and Vue.js applications, REST APIs, dashboards and real-time data.
Jul 2019 — Aug 2020
Visiro Global · Sri Lanka
Laravel and MySQL applications, API design, debugging and feature planning.
2017 — 2019
University of Bolton, UK
2025
University of California, Irvine
In progress
Amazon Web Services
Backend-heavy Laravel work: new systems, APIs, payment integrations, multi-tenant SaaS, admin panels and ERPs, performance fixes on slow applications, and technical leadership for teams that need a senior engineer.
Yes. I work remotely with clients around the world from Colombo (UTC+5:30). I work asynchronously with clear written updates, and we agree on call times that suit us both.
Hourly or fixed price. Fixed price fits work with a clear scope, usually after a short discovery phase. Hourly fits ongoing or evolving work. We talk rates once I understand the project.
Yes. I start by reading the code, mapping the architecture and listing the risks. Then we agree on what to fix first before adding features on top.
Yes: Vue.js, Inertia.js, Livewire, Alpine.js, Blade and Tailwind CSS. Backend systems and architecture are my core strength, though.
Send me a short description of your project on WhatsApp or by email. I’ll reply with questions, then we have a call. You get a written scope with diagrams and an estimate before any code is written.
Message me directly. No forms, no agency in between. You talk to the engineer who will do the work.
Me · Colombo, Sri Lanka
--:--Asia/Colombo · UTC+5:30
You
--:--Your local time
Remote by default Clear written updates, and calls at times that suit us both.
Colombo is UTC+5:30.