Zeeshan Ahmad

Most of the 25+ apps I have shipped run on one codebase across both stores. That is the whole argument for cross-platform: you pay for one build, one test suite and one release process instead of two.

Flutter is my default after five years in it, and I reach for React Native when the client already has a JavaScript team or a Supabase backend. I will tell you which one fits before we start, and why, rather than defaulting to whichever I feel like writing.

Where it does not fit, I say so. Heavy background location, deep OS integration and anything that lives or dies on platform SDK release day are still better native.

Lieferking was three Flutter apps on one .NET backend: consumer ordering, a restaurant partner app, and an in-store POS terminal. Tezz Delivery was a customer app plus driver dispatch. GoCargo was a consumer app plus a warehouse web portal. Freightly is the same pattern again, multi-tenant this time.

The hard part is never the screens. It is deciding where a permission is enforced, what a role can see, and what happens when two roles act on the same record at once. I have made those calls wrong before and I know which ones cost you later.

You own the repo from day one. I work on your GitHub, in branches and pull requests, so there is no handover event at the end.

I am an engineer, not a designer. If you bring a designer, I will build what they drew and push back where it will not survive a 320px screen or a 40-character German word. If you do not have one, I can work from references and ship something plain that functions, which is usually the right call for an MVP.

What I do own is the parts users actually feel: empty states, loading and error states, offline behaviour, form validation, and dark mode. Those are where most apps look unfinished, and they are almost never in the Figma file.

Pennywise is the honest example. The React Native version's UI came from Claude-generated mockups with no designer involved. It is on the Play Store, and the case study says exactly that.

Performance work starts with a profile, not an opinion. Flutter DevTools for jank and rebuild counts, the size analyser for the bundle, Firebase Performance for what real devices see rather than what my machine sees.

The usual culprits are the same every time: a widget rebuilding the whole tree on every keystroke, images decoded at full resolution into a 100px thumbnail, and a list that builds all 2,000 rows instead of the twelve on screen. FlashList and lazy builders fix the third one in an afternoon.

I also do offline-first work, which is a performance problem wearing a different hat. Pennywise runs entirely on-device with no login and no network call, and Firebase-backed apps like BoRoom and GoCargo cache so they still open on a bad connection.

I have shipped against Firebase, Supabase, .NET REST APIs and GraphQL. If you already have a backend and a backend engineer, I work to their contract and negotiate the shape of it up front. On Lieferking that meant settling the address payload with the backend engineer before either side wrote code, which is cheaper than discovering the mismatch at integration.

If you do not have a backend, I will build one. Firebase when you need auth, push and a document store tomorrow. Supabase when you want real Postgres, row-level security and to leave later without a rewrite.

Also in scope: Stripe and in-app purchases via RevenueCat, Google Maps and Places, Algolia search, FCM push, and OAuth. Each of those is live in something on the portfolio below.

I have taken 25+ apps through both stores, including the rejections. Store deployment is included in the project rather than billed separately: signing, provisioning, listings, screenshots, privacy declarations, and the back-and-forth with review when Apple wants something reworded.

After launch you get three months of free bug support. Beyond that, maintenance is real work and I price it as such, because every year brings a forced target SDK bump, a Flutter or React Native major, and at least one policy change that will pull your app down if nobody is watching.

Two of my longest relationships are maintenance, not builds: Parenting Hacks and a prank app at 10K+ downloads, both taken over from someone else and extended over years.