Software engineering studio
Nexus 22 Solutions designs, builds and runs production software — mobile apps, web platforms, APIs and the infrastructure underneath them. We do it for our own products, and we do it for businesses who need one team to own the whole stack.
FIG. 01 — ONE PLATFORM, EVERY SURFACE
How we work
Most studios only ever build for other people. We do both — and running our own platforms in production is what makes us useful on yours.
OUR SIDE
We fund, design, ship and operate our own software. That means we carry the pager, pay the cloud bill, and live with every architectural shortcut we take. There is no hand-off at launch — the thing has to keep working on a Sunday night.
YOUR SIDE
Small and mid-sized businesses get the same team and the same standard. No junior bench, no ticket queue, no six-week discovery phase before anything exists. You get engineers who have shipped this before, working directly with you.
Our products
These are ours — funded, built and operated in house. Detailed project specs and technical write-ups are being published section by section.
Live in production
A map-first venue and events platform for Southeast Asia.
HonestLee is one platform wearing six faces: an installable consumer web app, a native Android app, an admin console, a field-operations app for people on the ground, a backend API, and a media pipeline that resizes and serves every image at the edge. They all share one database and one data model — which is exactly the problem most multi-surface products fail to solve.
Under it: geospatial search over PostGIS, live maps, offline-tolerant mobile clients, real-time table service over websockets, multi-language digital menus, and an AI pipeline that extracts structured data from photographs.
Internal · in use daily
The tooling we use to run AI engineering agents against real codebases — task boards, structured signals, session state and audit trails, backed by Postgres. It is how we keep multiple agents working on one repository without them standing on each other.
Live
Self-hosted speed-test servers and a mobile measurement client that records real connection quality at a physical location — not a lab number. Multi-stream throughput, single-stream websocket tests, and the pipeline that stores and ranks the results.
In development
A dedicated project section with full technical specifications, architecture diagrams and build notes for each product is being written now. If you want the detail before it is published, ask us for it directly.
Request the specsCapabilities
We are not a front-end shop that sub-contracts the backend, or a backend shop that outsources design. The same people take a product from the first sketch to the server it runs on.
Native Android and iOS, and installable web apps that keep working when the network does not. We choose cross-platform where it saves you money and go native where the hardware actually matters.
Customer portals, booking flows, internal dashboards and admin consoles that stay fast when the table hits a million rows. Built as real applications, not a page builder with a login on top.
The API every one of your surfaces shares. Typed end to end, versioned, documented, and instrumented from the first commit so you can see what is slow before a customer tells you.
Deployed close to your users, on infrastructure that does not bankrupt you as you grow. Global CDN, edge compute, object storage and image pipelines — with the monthly bill as a design constraint.
Extraction, enrichment, semantic search and recommendations. We put hard cost controls and human review around every model call, because an AI feature that quietly bills per record is not a feature.
Live maps with thousands of points that still pan smoothly, boundary and territory logic, proximity search, routing and check-ins. Geospatial is its own discipline and we have shipped it at scale.
The stack
Every product we build has the same shape. Knowing where a problem lives is most of fixing it — so we keep the boundaries honest.
The apps and screens themselves. Fast first paint, sensible offline behaviour, accessible by default, and a design system so the tenth screen looks like the first.
CDN, edge functions, caching rules and an image pipeline that serves the right size to the right device instead of shipping a 7 MB photo to a phone.
The API, auth, permissions, background jobs, webhooks and realtime channels. Written once and shared by every surface, so a rule change ships everywhere at once.
A relational schema designed to still make sense in three years, with migrations, backups, spatial indexes and vector search where it earns its place.
For small & mid-sized business
Most SME software projects fail the same way: a big agency bills a discovery phase, a junior team writes it, and eighteen months later nobody can change anything. We work the opposite way round.
START HERE
Two to three weeks. We take a fuzzy idea and turn it into an architecture, a scoped build plan with real numbers, and a clickable prototype you can put in front of customers.
MOST COMMON
The real thing, shipped. A production application with the infrastructure, the admin tooling and the deployment pipeline around it — priced as a fixed scope so you are not watching a meter.
ONGOING
A monthly retainer for teams who would rather not hire an engineering department: we keep it patched, watched and moving forward at an agreed pace.
Not sure which one you need? Tell us what you are trying to do and we will say so plainly — including if the honest answer is that you should buy something off the shelf instead.
Process
This is a real sequence, so it is numbered. Each step ends with something you can hold — not a status update.
We ask the awkward questions early: who uses this, what breaks if it is late, what already exists, and what you are actually measured on. You end this step with a written scope and a fixed price.
OUTPUT — scope document · fixed quote
Data model, API surface, infrastructure and the cost curve, decided before code. This is the step most projects skip and every one of them regrets.
OUTPUT — architecture & schema · running cost estimate
Two-week cycles, each ending with something deployed you can open and use. You get a staging environment from the first week — progress is visible, not reported.
OUTPUT — deployed increments · staging access
Launch, monitoring, backups and a handover that assumes we will not be here forever. Stay on a retainer or take it in house — both are fine, and the documentation is the same either way.
OUTPUT — production launch · docs · handover
How we think
Repository, cloud accounts, domains, data, credentials. If you decide to leave, you leave with a working system — not a negotiation.
Infrastructure cost gets designed at the same time as the schema. A feature that triples your monthly spend is a feature we will argue with you about first.
Postgres, typed languages, plain HTTP. We save the novelty for the parts of your product that are genuinely novel.
You talk to the people writing the code. No account manager relaying your question to someone in another timezone.
A staging URL in week one tells you more than a hundred-page requirements document, and it costs less to change.
If off-the-shelf software solves it, or the project is not viable, we will say so before you spend the money. It costs us a contract and earns the next one.
Questions
It depends on scope, but we quote a fixed price before any build starts, so you are never watching an hourly meter. A product sprint is a small, defined engagement; a first build is quoted from the scope document it produces. If your budget and the scope do not meet, we will tell you during the first conversation rather than three months in.
You will have a staging environment you can open within the first two weeks of a build. A focused first version of a real product typically reaches production in two to four months, depending on integrations and how quickly decisions come back from your side.
Yes, completely, and from day one — not on final payment. Work happens in your repository under your organisation, deployed to cloud accounts in your name. There is no proprietary platform layer you would have to buy your way out of.
Often, yes. We start with a paid technical audit: what is there, what is safe, what it will cost to maintain versus replace. Sometimes the honest recommendation is to keep what you have and fix three specific things — we will say that if it is true.
The same senior engineers who scope the work. We are deliberately a small studio and we do not run a junior bench or sub-contract the parts we find boring. If we cannot staff your project properly, we will tell you instead of taking it.
We use AI tooling in our own engineering workflow, under human review, the same way we use compilers and test suites — and we build AI features into products when they solve a real problem. Every line that ships is reviewed by an engineer who is accountable for it.
Nexus 22 Solutions LLC is registered in Sheridan, Wyoming, and we work with clients worldwide, remotely. Contracts are US-based. In practice the only thing that matters is overlap hours, and we will agree those before we start.
Retainers cancel with thirty days' notice. Fixed-scope builds can be stopped at any agreed milestone. In either case you keep everything built to that point, deployed and documented — that is the whole point of you owning the accounts.
Start a project
A few lines is enough to start. We will reply with what we think it takes, what it roughly costs, and whether we are the right people for it.