The four areas we work in

We are deliberately narrow. These are the areas where we have shipped enough to know where the problems hide.

01 AI Applications

Typical work: Turning an unstructured pile of calls, forms and scans into something queryable; automating a step a person currently does by hand.
Constraints we plan for: Bilingual input, messy source data, outputs that need a human in the loop, cost per document at volume.
What you get: An automated step that fails loudly, not silently.

AI Applications

Models wired into systems that already exist

Most of the effort goes into the pipeline around the model, not the model itself: pulling clean input out of source systems that were never designed to hand it over, handling the cases the model gets wrong, and noticing when the output has drifted.

Speech and document pipelines, Arabic and English
Retrieval and extraction over existing databases and file stores
Evaluation and monitoring, so a quality regression is visible before a customer finds it

02 Enterprise Hosting

Typical work: Standing up or taking over the infrastructure a product runs on; untangling a delivery pipeline; replacing an access model that has accumulated exceptions.
Constraints we plan for: Data-residency rules, constrained bandwidth and mobile networks, legacy systems that cannot be replaced, wide device spread.
What you get: Infrastructure somebody can explain, and logs that answer questions.

Enterprise Hosting

We run what we build

Hosting is where the promises get tested. A transcode ladder that does not match the audience’s bandwidth, a CDN caching the wrong manifest, a permissive default nobody reviewed: these are infrastructure problems, and they stay ours. We deploy on our own infrastructure or on yours, including where the data is not permitted to leave the country.

On-premise, private-cloud and hybrid deployment, documented so your team can run it
Video and audio delivery end to end: ingest, transcode ladders, packaging, CDN and origin
Authentication, authorisation and audit trails designed to be queried, not merely written

03 SaaS & Data Platforms

Typical work: Taking a product that serves one customer and making it serve a hundred; making releases safe to ship to all of them at once; turning a stream of events into something a person can act on.
Constraints we plan for: Tenants on different plans and versions at once, data that must not cross between them, usage that spikes without warning, retention cost at volume.
What you get: A product you can onboard a customer onto on a Tuesday afternoon, and numbers you can act on.

SaaS & Data Platforms

Software you sell as a service, and the numbers behind it

Serving one customer is not serving a hundred. Tenants have to stay isolated from each other, plans and roles have to hold, onboarding has to work without you in the room, and a release reaches everyone at once so it needs a way back when it goes wrong. The other half is what the product produces: events that arrive complete, are kept long enough to compare against, and roll up into something worth looking at.

Multi-tenant architecture: tenant isolation, roles and permissions, self-serve onboarding
Subscriptions, metering and billing, with staged releases, health gates and rollback
Event ingestion, retention tiering, and the dashboards and alerting built on it

04 Tech Sponsorships

What we provide: Hosting, email on your own domain, and the tooling around them. Sometimes an afternoon of engineering time, when that is what is actually blocking you.
Who it is for: Startups before revenue, students, and small teams with something built or nearly built. Iraq and Europe both.
What you get: Infrastructure you do not pay for, and no claim on what you build.

Tech Sponsorships

We cover the running costs for work worth running

Somebody covered ours once. So we keep a share of what we run for people who have built something and cannot yet pay to keep it online. There is no application form and no committee. Send us a paragraph about what you have made and what it needs, and we will tell you quickly whether we can carry it.

Hosting and email, on infrastructure we already operate
No equity, no invoice, and no obligation to credit us
A straight answer within a few days, either way

We know both markets from the inside

Nisaba was the Sumerian goddess of writing and record-keeping, the one who kept the accounts straight. That is roughly the job.

Why we go deep instead of wide

We have spent over ten years on production software in Germany and Iraq. In Europe that meant large projects for companies where a bad release is measured in revenue per minute. In Iraq it meant building the systems themselves, at national scale, under constraints most vendors never have to design around. Nobody reads that market from a slide deck; we built inside it, and very few teams know it the way we do.

That depth is the reason we specialise rather than spread. Every engagement is staffed by senior engineers who have shipped this kind of system before, and the people who scope the work are the people who build it; nothing is won by one team and handed to another.

NISABA, in short

Team: Senior throughout, hands-on from scoping to handover.
Where: Munich, Germany · Baghdad, Iraq.
Markets: Iraq · Germany and the rest of Europe.
Experience: 10+ years shipping software that other people depend on.
Domains: AI applications, enterprise hosting, SaaS and data platforms.
Working languages: English and Arabic.

How we work

We say what we do not know

If a problem sits outside what we have actually shipped, we will tell you and point you elsewhere. Learning on a client’s budget is a bad trade for both sides.

The people you meet do the work

Whoever you speak with is writing the code and answering for it later. Nothing is scoped by one team and handed to another once the work is won.

Built to be handed over

Documented, tested and deployable by your own team. You should not need us on retainer to change a configuration value.

What we have built

Two of these are our own products. Two are client work we can describe but not name. The last is how all of them get designed.

01 · Our product

NISABA EXPO

An exhibition management system with a CMS behind it: exhibitors, stands, schedules and the public site, all edited by the organiser rather than by us. Built for the day the doors open, when everything changes at once.

CMS Exhibitor portal Scheduling Multi-tenant

02 · Client work

Streaming that holds under load

Live and on-demand video to subscribers over networks where the available bandwidth moves hour to hour. Ingest, transcode ladders, packaging, CDN and origin, with playback telemetry so a bad session traces to a segment and an edge instead of being argued about.

HLS / DASH Transcode ladder CDN & origin Playback telemetry

03 · Our product

Kertas

A platform for teaching that has moved online: course material, assignments and marking in Arabic and English, built to stay usable on the connections students actually have.

E-learning Arabic & English Coursework & marking Low-bandwidth

04 · Client work

Enterprise AI CRM

CRM systems where the model does real work: summarising calls and correspondence, drafting replies, and surfacing the record someone needs, wired into the systems a company already runs. With an evaluation set running, so a quality regression shows up before a customer finds it.

LLM pipelines CRM integration Arabic & English Eval harness

05 · Design

Interfaces built for the AI age

A chat box is not an interface for everything. When a model does part of the work, the screen has to show what it did, how sure it is, and where a person can override it, in two languages and both text directions. That is the design brief on everything above.

Design systems Human in the loop Bilingual & RTL Accessibility

Two of these are ours and we can show them. For the client work we can go further in a conversation than we can in public: scale, architecture, what went wrong.

Tell us what you are building

A paragraph is enough to start. If it is not something we can help with, we will say so quickly instead of booking a discovery call about it.

How to reach us

Mail reaches the whole team. We are in Munich and Baghdad, so between us there is usually someone at a desk.

Where we are

Munich, Germany · Baghdad, Iraq
Email Address

info@nisaba.uk
Telephone (Munich)

+49 152 36968148
Telephone (Baghdad)

+964 771 559 4158

We read everything that arrives and reply to anything with a real question in it.

Send a message

Full Name *

Please provide your full name.

Email *

Please enter a valid email address.

Company *

Please tell us where you work.

Closest area

What you are working on *

A sentence or two about the work, please.

Questions

The things people ask before they get in touch. If yours is not here, ask it directly. We would rather answer than have you guess.

AI applications, enterprise hosting, and SaaS and data platforms. We stay inside those areas and do not claim everything. They are where we have shipped enough to know where the problems hide. Separately, we sponsor hosting for a handful of startups and students each year.

We tell you before we start. For work that needs a larger team we will say so up front, and where it helps we will introduce you to people we have worked with. Taking on something we cannot staff would waste your time and ours.

More than most, because we built there rather than visited. National-scale systems, on infrastructure that is not always there, with procurement, connectivity and data-residency rules that no European reference architecture accounts for. The engineering standard we hold that work to came from years of large projects for companies in Germany and across Europe.

As a pipeline problem, not a model problem. Most of the effort goes into the input, into a person handling whatever the model is unsure about, and into an evaluation set that keeps running so quality regressions surface early. Arabic and English are both first-class.

Yes. On-premise, private cloud and hybrid deployments are normal for us, and we have built for cases where data is not permitted to leave the country. We will also hand over the deployment so your team can run it without us.

You do. We write it to be handed over: documented, tested and deployable by your own engineers. If you never need to call us again after the engagement ends, that is a success rather than a lost account.

With a conversation about what you are building and where it is straining. If it looks like a fit, we scope a first piece of work small enough to be judged on its own before anyone commits to more. We do not quote a timeline before we understand the problem.

Something not covered?

Ask us directly. Technical questions get technical answers, not a callback from someone who will have to check.

Writing Tomorrow’s Code, Since Day One

What we do