AI Development
Agents that do real work against your data, not a chat box bolted to a landing page.
Founder & CEO · ColorPark
One engineer who scopes the problem, writes the code, ships it to production, and stays accountable for how it runs. You work with the person writing the code, not an account manager relaying status.
Open — 1 project slot
Founder-led
The person estimating your build is the person who will write it. Estimates come back tighter, because nobody is guessing on someone else's behalf.
A schema change, a provider swap, a scope cut. You get an answer inside one business day, because there is no internal round trip to approve it.
Handover does not mean the knowledge walks out. I still maintain systems I shipped years ago, for clients who never had to re-explain them.
I don't ship AI demos. I ship systems that still run at 3 a.m. — monitored, maintained, and boring in the best possible way.
Capabilities
Most projects need three of these four at once. One person holding all four is why the seams between them stop being your problem.
Agents that do real work against your data, not a chat box bolted to a landing page.
Business systems that hold up under load, audit, and the third round of requirements.
Infrastructure you can rebuild from a repository, and a deploy nobody is afraid to run on a Friday.
We find what an attacker would find, then close it and prove it stayed closed.
Ventures
Software engineering and cloud delivery for global clients.
Founder & CEO
Creative design agency for brand, product UI, and marketing.
Founder & CEO
Penetration testing and managed security operations.
Founder & CEO
WordPress development and maintenance for content businesses.
Founder
Global digital commerce platform.
Deputy Head of Software Development · 2017 to present
How I work
We map the problem, the constraints, and what done looks like. You get a written scope before any code, so the estimate is something you can hold me to.
Data model, service boundaries, failure modes. Decisions get written down with the reasoning, so the next engineer inherits context instead of archaeology.
Small releases behind CI, into a real environment you can click through. You see progress every week, not in one reveal at the end.
Monitoring, alerts, runbooks, and a person who answers. A system is not finished when it deploys. It is finished when it runs without you watching it.
That is what the first two weeks look like. Book a call to scope yours.
Engagement
The same engineer in all three. What changes is the commitment, the cadence, and how much of the running I keep afterwards.
A defined outcome with a written scope, a fixed shape, and a delivery date. Best when you know what you need and want it built once, properly.
Best for new builds, migrations, and replatforms.
Scope a buildA standing block of my time each month for the systems you already depend on. Features, fixes, upgrades, and the on-call attention that keeps small problems small.
Best for live products with a roadmap and real users.
Discuss a retainerA short, hard review of a system you already run: architecture, security, cost, and what breaks first. You get a written plan your team can execute without me.
Best for teams with engineers but no direction.
Book an advisory sprintStack
Credentials
Dates, institutions and titles, stated plainly so you can check them before the call rather than after it.
Daffodil International University
ElectronicFirst.com FZ LLC, a global digital commerce platform
Ramlit Limited, ColorPark.io, and xCyberSecurity.io
Remote-first, working with clients worldwide
Questions
I work from Dhaka, UTC+6, and keep deliberate overlap with Europe and the US East Coast. Expect a reply within one business day, and calls booked in your working hours rather than mine.
Project builds start at a scope worth writing down, which in practice means two weeks or more. Retainers run monthly with one month of notice either way. An advisory sprint is the smallest way in, and it is the most common first step. I hold one project slot at a time, so the availability line at the top of this page is a real constraint rather than decoration.
I do. On larger builds I bring in engineers from Ramlit Limited, and I stay the technical owner: I review every merge, and I remain the person you talk to.
Your choice. I hand over documentation and runbooks and step back, or I stay on retainer and operate it. Either way, nothing goes live without monitoring and a tested way to roll it back.
You do, on final payment: repositories, infrastructure definitions, and credentials. Handover is a scheduled task with a checklist, not an email on the last day.
Send a short note about the problem and the deadline. If it is a fit, we book a call, and you get a written scope before you commit to anything.
Still unsure whether it is a fit? Book a call and ask.
One call, no pitch deck. If I am not the right person for the job, I will say so on the call and point you somewhere better.
Prefer to type? Message +880 1723 741224 and I will reply within one business day.
Search across services, blog posts, portfolio & freelance services