Product Development Head at Nexsys-One
Since 2022 I’ve led product development at Nexsys-One in Dubai, running the web, mobile and performance teams behind its cloud platform for telecom and renewable energy operators, and bringing AI into it where it earns its place. Before that I spent six years as a senior engineer and team lead on Asfar Group’s travel booking platforms, and five in Lahore writing server-side PHP, the last of them leading a team.
01Interactive demo
Legacy systems rarely survive a rewrite in one go. They’re taken apart a piece at a time, with users moved over gradually and every step checked. Here an AI agent plans the order and does the work, and you decide how much it may do alone. Click any module to move it in one go and see what happens.
Agent proposal
Runs in your browser. A made-up application, not based on any company’s software or data.
02Expertise
I design and run platforms on AWS: EC2, S3, RDS, Lambda, API Gateway and CloudFront, containerised on ECS and EKS. Deployments go through CodePipeline, Jenkins or GitHub Actions, and the infrastructure lives in CloudFormation templates so it can be rebuilt.
I bring AI into existing systems where it pays off. In telecom operations that means predictive maintenance, anomaly detection, network optimisation and digital twin simulations, with agents that triage alerts and handle routine decisions. Every agent has a person to hand over to when it isn’t sure.
I’ve looked after MySQL databases well past a terabyte. That means replication, sharding and partitioning to keep them available, EXPLAIN and Performance Schema to keep queries fast, and backup and recovery plans for the day something fails.
I lead web, mobile and performance engineering teams, from hiring through to delivery. A good part of my week goes on keeping the technical work tied to what the business, the product owners and our clients actually need.
03Selected work
Telecom infrastructure management
A cloud platform that brings the day-to-day running of telecom and renewable energy sites into one place for operators, with predictive maintenance and intelligent alerting built in. I lead its product development and the engineering teams that build and run it.
Travel booking engine
A booking platform for flights, hotels, car hire and insurance, with payment gateways built in. Partners run white-label versions of it under their own brands.
Yacht booking and ERP
Yacht booking connected to an ERP, so scheduling, billing and customer self-service work as one system rather than three.
Startup and investor network
Secure APIs and a backend built to scale for a fast-growing platform that connects founders with investors.
Investment marketplace
A marketplace with collaboration tools for investors and stakeholders, and Elasticsearch-powered search that filters by location.
Ideas and debate platform
A Drupal crowdsourcing platform that collects and evaluates people’s ideas in real time, integrated with Salesforce and PayPal.
Employee engagement
An enterprise platform for internal communications, peer recognition and employee perks, which I contributed to.
Family memories app
A secure cloud app where families store and privately share photos and moments from their children’s lives, with media upload, content filters and sharing controls.
Points-based customer rewards
A mobile-friendly rewards system where customers earn and redeem points, with partner integrations.
04Experience
2022 — NowDubai, UAE
Nexsys-One · Telecom software
I lead product development and the engineering teams behind it: cloud-native architecture on AWS with Docker, Kubernetes and automated CI/CD, terabyte-scale MySQL, and the AI behind its predictive analytics, real-time monitoring and digital twin simulations. I manage and mentor the web, mobile and performance teams, and work with executives and clients to align the roadmap with business goals.
2016 — 2022Dubai, UAE
Asfar Group · Travel and tourism
I led the development of Asfar’s travel products, including AsfarTrip, from requirements through to long-term support. I mentored the team through code reviews and complex work, improved database performance, APIs and caching for systems under heavy booking load, ran sprint planning and the backlog, and gave partners the API documentation and sandbox environments they needed to integrate.
2014 — 2016Lahore, Pakistan
PutItOut · Software design and development
My first lead role. I coached and mentored the team, set up our quality practices, turned user requirements into estimates and plans, took part in design and code reviews, and interviewed candidates.
2012 — 2014Lahore, Pakistan
Coeus Solutions · Software design and development
I built and maintained web applications in PHP and Laravel, including MySQL schema design and query optimisation, turned stakeholder requirements into working features, and helped resolve issues for clients.
2011 — 2012Lahore, Pakistan
HBS Technologies · Remittance software
Where I started. I developed modules for an in-house international remittance platform, led internal development projects, built features based on market research, and fixed customers’ live-site issues.
05Approach
The request a project starts with is usually a symptom. Before designing anything I look for what’s behind it: a queue that keeps backing up, a query that got slow, a process people quietly work around. I don’t start building until I can describe that problem in one sentence and the people who own the system agree with it.
Most technical decisions are easy to change later. A few aren’t, like how the data is structured, where state lives and which parts are allowed to be slow. I pull those out at the start, work out what each option would cost, and decide them with the team. Everything else I leave open until we know more.
Each agent gets a clear task, a fixed set of tools, an output we can check, and a point where it has to stop and ask a person. We write the tests before adding capability, and an agent only gets more freedom once its track record justifies it.
Teams end up shaped like the systems they look after, so I plan the two together. Each team owns a clear area and works to the same quality bar. I spend as much time on how teams hand work to each other as on how the services talk to each other.
We load-test before launch and practise the likely failures instead of waiting for them. The last part of the job is the handover: monitoring, runbooks, and someone on your team who understands the decisions well enough to defend them without me.
Cloud, data, AI and the people who run them, working as one system.
06Principles
Age has little to do with it. A new codebase that nobody on the team can describe end to end is already legacy. That’s why the first thing I produce on a project is usually a diagram. Until the people who own a system can see its shape, every estimate is a guess.
Agents tend to fail where their scope was assumed instead of agreed. Give an agent a defined task, a fixed set of tools, an output you can check and a point where it must stop and ask. Widen its freedom as its track record grows, and narrow it again when things go wrong. Testing has to come first. If it’s left for later, it rarely happens.
Every platform gets rewritten eventually. What matters is whether it happens on a date you chose. Plan it ahead (which component first, at what load, behind which interface) and it’s just another project. Wait until something breaks and it becomes the year and a half nobody wants to talk about.
Features decide what a system does. How it fails decides what it costs to keep running three years later: the 3 a.m. pages, the manual fixes, the migrations nobody wants to touch. An architecture review that only looks at the happy path is reviewing the cheapest part.
07Services
For executives and boards
Sessions for the decisions you can’t delegate: build or buy, platform or product, which system to retire first. We meet when a decision is coming up, not on a fixed schedule.
Ask about thisBefore you build
You bring an idea or a draft design, and we work out its structure together: what the parts are, how they depend on each other, what gets built first, and which failures it has to survive. We estimate cloud costs and agree the data model before anyone writes code.
Ask about thisFrom existing systems to AI agents
A larger engagement across your whole system: a review of what you run today, a plan for bringing AI agents into it, and a plan for the teams needed to run both. It works best with an executive sponsor who can approve retiring old systems.
Ask about this08Contact
Tell me what you’re running, what isn’t working and when it needs to change. A short paragraph is plenty.
Messages come straight to me and I reply in writing, usually with a first view of the problem and a few questions. If it looks like a fit, we’ll set up a call with whoever owns the system on your side, and I’ll follow up with a written scope.
Connect on LinkedIn