Skip to content
New ওয়েবে বাংলা টাইপোগ্রাফি — যে সমস্যাগুলো নিয়ে কেউ লেখে না

One idea in. A fully deployed system out.

Website, mobile app, staff and payroll, servers. One team builds all of it, and the same team is still accountable a year after launch.

One team
design, build and run
Written SLA
on every system we operate
Any size
one counter to one ministry

Built for the constraints of these sectors

  • Government
  • University & Education
  • Hospital & Healthcare
  • Bank & Finance
  • Automotive
  • Retail & Supershop
  • E-commerce & D2C
  • Private Enterprise

Why organizations call us

The building was never the hard part.

Four suppliers can each do their work correctly and still leave a system nobody owns. The designer signs off. The developer signs off. The server was configured once, a year ago. What falls between them becomes yours by default. These are the gaps we are asked to close most often.

  • Four suppliers, one system

    A designer, a developer, a server admin, and whoever holds the maintenance contract. Every one of them is doing their part correctly. The failures live in the gaps, and nobody is contracted to look there.

  • The project ends at launch

    Go-live is written as the final milestone, so it is also the last day anyone is responsible. Six months on there is an unapplied security patch and no one whose job it was.

  • A website you have to phone someone to edit

    Changing a price or posting a notice means raising a ticket and waiting two days. The information that most needs to be current is the information you control least.

  • Nobody owns the server

    It was configured once and has run ever since. Whether the backups exist, and whether they restore, are questions that get answered for the first time during an outage.

  • Operations live in parallel spreadsheets

    Attendance in one file, stock in another, payroll in a third, each kept by a different person and none of them agreeing. The system covers your customers and stops at the office door.

  • Growth breaks the architecture

    Built for one branch, one warehouse, one department. Opening the second requires a rewrite rather than a setting, and it lands in the month you have least time for one.

What we do

Seven capabilities. One accountable team.

Most engagements use three or four of these. We scope to what your organization actually runs on and quote the pieces you need, so nothing here is sold as a bundle.

Who we serve

Every sector has one constraint that decides everything.

For a hospital it is privacy. For a university it is admission week. For a bank it is the first time a customer disputes a transaction. The core architecture is shared. What sits on top of it is built for the constraint.

  • Government

    Multi-department coordination, data residency rules and public-facing transparency, all binding at the same time.

    A multi-department portal with audit-trail logging throughout. Citizen services and internal workflow run in separate layers over one shared backbone.

    • Role-based multi-department access
    • Full audit trail logging
    • Citizen-facing service portal
    • Data localization compliance
    The GovTech practice
  • University & Education

    Admission week is the traffic peak of the year. Result publication is the one release that cannot slip.

    One academic platform covering admission and payment, student records, results and LMS integration, provisioned for the spike rather than the average.

    • Online admission and payment
    • Result and transcript management
    • LMS and course integration
    • Parent and guardian access layer
  • Hospital & Healthcare

    Patient privacy, double-booked appointments and shift coverage, in the setting where the margin for error is smallest.

    Encrypted patient records behind strict access control, appointment booking with automated reminders, and staff scheduling in one system rather than three.

    • Encrypted patient records
    • Appointment and queue management
    • Staff shift scheduling
    • Privacy compliance controls
  • Bank & Finance

    Transaction security, regulatory reporting and fraud handling, where one visible mistake costs the trust of everyone watching.

    End-to-end encrypted transaction flows, a compliance-ready audit dashboard, and fraud alerts that reach the customer inside the session rather than in a next-day email.

    • End-to-end encrypted transactions
    • Compliance-ready audit dashboard
    • Real-time fraud alerts
    • Multi-factor authentication
  • Automotive

    Stock spread across branches, service slots booked by phone, and dealership performance nobody can see in one place.

    Centralized inventory and dealership management, with a customer-facing booking platform on the same data, so the slot a customer picks is one the workshop actually has.

    • Multi-branch inventory sync
    • Online service booking
    • Dealership performance dashboard
    • Customer service history log
  • Retail & Supershop

    Billing speed at the counter, stock accuracy across branches, and branch-wise sales visibility, all needed at once.

    Barcode POS billing with live multi-branch stock sync, reporting into one central dashboard, and offline billing so the counter never stops.

    • Barcode POS billing
    • Multi-branch stock sync
    • Supplier and purchase management
    • Branch-wise sales analytics
  • E-commerce & D2C

    A campaign spike, a payment gateway that fails at the worst possible moment, and delivery updates that never reach the customer.

    A storefront provisioned for peak load, several payment gateways so one failing does not mean checkout failing, and courier-integrated tracking from order through to the door.

    • High-traffic-ready storefront
    • Multi-gateway payment
    • Courier and delivery tracking
    • Abandoned cart recovery
  • Private Enterprise

    Workflows that match no off-the-shelf product, and a scaling curve nobody can predict on the day of purchase.

    A custom ERP-lite core with a client and vendor portal, built so a second branch or a new department is a configuration change rather than a rebuild.

    • Custom workflow automation
    • Client and vendor portal
    • E-commerce integration
    • Scalable multi-branch support

Government & public sector

Technology for a more connected government.

UXAtom designs and develops secure, scalable software platforms that help government organizations modernize operations, automate workflows, manage people, and deliver better digital services.

  • National ministries
  • Government agencies
  • Directorates & departments
  • State & provincial government
  • Municipal & local government
  • Public health systems
  • Education authorities
  • Regulators & public bodies
  • State-owned enterprises

Process

Six steps. The last two have no end date.

Every project runs the same sequence, so you always know what is happening now and what comes next. Deploy is step four of six.

  1. 01

    Discover

    Week 1–2

    We interview the people who will use the system every day, not only the people commissioning it, and audit whatever is running today. The goal is to find the real constraint before anyone opens a design tool.

    • Stakeholder interviews
    • Current-system audit
    • Scope document
  2. 02

    Design

    Week 2–6

    Research, information architecture and interface design, with the system architecture planned in the same weeks rather than after it. A screen the data model cannot serve gets caught here, while changing it is still free.

    • Flows and wireframes
    • High-fidelity UI
    • System architecture
  3. 03

    Develop

    Week 5–14

    Sprints with a review at the end of each one, so the first time you see the system is not the week it is due. Accessibility and performance are checked automatically on every commit rather than audited at the end.

    • Working software each sprint
    • API and schema
    • Test coverage
  4. 04

    Deploy

    Week 12–16

    Production launch with a zero-downtime strategy, and with backups and monitoring provisioned as part of go-live. Where an old system is being replaced, the two run side by side until the new one has earned the switch.

    • Production environment
    • CI/CD pipeline
    • Backup and rollback plan
  5. 05

    Manage

    Ongoing

    The subscription phase. Monitoring, security patching and bug fixes against a tiered SLA, through a support channel that reaches someone who knows your system rather than a queue.

    • SLA response times
    • Monthly security patching
    • Quarterly health report
  6. 06

    Scale

    Ongoing

    Feature and infrastructure expansion as the business grows: a second branch, a new department, a traffic profile nobody forecast. All of it planned against the same architecture, so growth stays a configuration change.

    • Capacity planning
    • Feature request pipeline
    • Infrastructure expansion

Design system

This site runs on the system we build for clients.

Everything you are looking at resolves through the same token layer and the same component contract we hand over at the end of an engagement. Open the inspector. There are no one-off values in here to find.

309
Design tokens
31
UI primitives
9
Stylesheet layers
AA
WCAG 2.2 target
Explore the system
theme.css Tailwind v4
@theme inline {
  --color-primary: var(--primary);
  --color-surface:  var(--surface);
  --color-border:   var(--border);
}

@utility ux-card {
  background-color: var(--surface);
  border: 1px solid var(--border);
  border-radius: var(--radius-xl);
  box-shadow: var(--inset-shadow-highlight);
}

Tokens, not values

Every colour, space and duration resolves through a semantic token. Rebranding is a variable change, not a refactor.

Variants as contract

Each component declares its variants, sizes and states in one table. If it is not in the table, it does not exist.

Accessible by construction

Focus rings, ARIA wiring and reduced-motion handling live in the primitive, so no consumer can forget them.

One source, two surfaces

The Figma library and the code library are generated from the same token file, so they cannot drift apart.

Learn

The clearest sample of our work is free to read.

We build and maintain a documentation platform in the open, on the same components, the same tokens and the same accessibility and performance standards we hand to clients. If you want to know how we structure information and write for a technical audience, this is the evidence.

6
Courses
12
Lessons
24
Books & PDFs
2
Languages
  • Long-form technical writing

    Kubernetes networking, system design, the JavaScript event loop. Written for people who can already do the job and want to know why it behaves the way it does.

  • Quizzes that mark themselves

    Each topic ends with questions and explanations, so reading turns into checking whether you actually understood it.

  • Built to be taken offline

    Every topic has a print view that turns the whole thing into a clean PDF through the browser. No account, no export queue.

  • Written in Bangla, not translated

    The Bangla track is its own body of work, with its own topics and its own typography, including the freelancing and career material English tutorials skip.

Tech stack

Tools we are fluent in, not just familiar with.

Design tools, front-end frameworks, the services behind them and the infrastructure they run on. We pick what fits your team, then work in it at a senior level. Nobody learns a framework on your budget.

  • Figma

    Design, variables and libraries

  • Storybook

    Component documentation

  • Framer

    High-fidelity prototypes

  • Vue

    Composition API, script setup

  • Nuxt

    SSR, hybrid rendering, SEO

  • React

    When your team already runs it

  • TypeScript

    Strict mode, no implicit any

  • Tailwind CSS

    v4 tokens and theme layer

  • Motion

    Interface animation

  • Node.js

    Nitro and edge runtimes

  • PostgreSQL

    Primary datastore

  • Supabase

    Auth, storage, realtime

  • Vercel

    Preview deploys and edge

  • Cloudflare

    CDN, Workers, R2

  • GitHub

    CI, budgets, review

  • Anthropic

    Claude for product features

  • LangGraph

    Agent orchestration

  • pgvector

    Retrieval and embeddings

Writing

Notes from inside the work.

Patterns we keep reaching for, mistakes worth avoiding, and the occasional teardown.

design9 min read

ওয়েবে বাংলা টাইপোগ্রাফি — যে সমস্যাগুলো নিয়ে কেউ লেখে না

লাইন-হাইট, মাত্রার সংঘর্ষ, uppercase-এর অর্থহীনতা আর ফন্ট ফলব্যাক। ল্যাটিনের জন্য বানানো একটা ডিজাইন সিস্টেমে বাংলা বসাতে গিয়ে যা যা ঠিক করতে হয়েছিল।

তাসনিম আরা

তাসনিম আরা

প্রোডাক্ট ডিজাইনার

design-systems8 min read

ডিজাইন সিস্টেম বানানোর আগে যে ভুলগুলো এড়ানো দরকার

বেশিরভাগ টোকেন সিস্টেম প্রথম রিব্র্যান্ডেই ভেঙে পড়ে। কারণটা প্রায় সবসময় এক — দুইটা লেয়ার আছে, দরকার ছিল তিনটা।

নাফিস রহমান

নাফিস রহমান

ডিজাইন সিস্টেম লিড

design-systems7 min read

Design tokens that survive a rebrand

Most token systems break the first time the brand changes. The fix is a layer most teams skip — and it costs about a day to add.

Elena Marsh

Elena Marsh

Design Systems Lead

Engagements

Three ways to work together.

Every engagement is fixed-fee and quoted once we understand the problem. No hourly billing, and no change-order theatre. Scope changes get re-quoted in writing before anyone starts on them.

Design Sprint

A focused answer to one hard question.

Timeline
2 weeks
Best for
Validating a concept before committing a roadmap to it
  • Discovery and stakeholder interviews
  • Competitive teardown
  • Concept directions
  • Clickable prototype

Product Engagement

Most chosen

The full build, from research through to shipped product.

Timeline
8–14 weeks
Best for
Teams launching or rebuilding a core product surface
  • Everything in Design Sprint
  • End-to-end UX and UI
  • Design system and token architecture
  • Production front-end in your stack

Embedded Partner

A senior team inside yours, month to month.

Timeline
Rolling, 3-month minimum
Best for
Scale-ups that need continuous design and engineering capacity
  • Dedicated designer and engineers across front-end and backend
  • In your Slack, standups and sprint planning
  • Design system ownership and governance
  • Quarterly roadmap and metrics review

A 30-minute call, no deck. Full IP transfer and a checkpoint you can stop at.

FAQ

Questions we get asked first.

If yours is not here, email us. We answer every one, usually the same day.

Both ends of the range, deliberately. A single-branch shop that needs a till and a stock system, and a government department or bank that needs a multi-department platform with a full audit trail. The scope and the compliance burden differ enormously. The engineering standard does not. We are a poor fit for a one-off brochure site with no system behind it, and we will usually say so on the first call.

Launch is step four of six. From go-live the system moves onto a maintenance subscription with a tiered SLA: monthly security patching, bug fixes, a named support channel and a quarterly health report. It is a paid arrangement rather than a goodwill promise, and that is exactly why a response time can be held against it.

Not contractually. The maintenance subscription is cancellable, and the code, the data and the infrastructure are yours. In practice any team taking over a system needs a handover period, so we write architecture notes and runbooks as we go rather than assembling them in the month somebody leaves. We would rather you renew because it is the obvious choice than because leaving is painful.

Yourselves. Every build ships with a control panel for content, pages and the settings that change often: a price, a public notice, a term date. Structural changes and new features still come to us, because those are code, and a page builder that lets anyone restructure a live system creates a different kind of problem.

That is the Consulting and Digital Transformation pillar, and it is the most common starting point for larger organizations. The old system keeps running while the new one is populated and tested, and the cutover happens in phases so there is always something to fall back to. We scope data migration as its own piece of work, because it is usually the longest part of the project and almost always the one that gets underestimated.

Not yet, and we would rather say so plainly than show you a stock screenshot. What is published is self-initiated concept work, labelled as such, plus this website and the design system it runs on. You can open and audit all of it without asking us for a reference. The day there is client work we are permitted to name, it will appear here with the client attached to it.

One useful email a month

Design system patterns, front-end techniques and case study breakdowns. No promotions, no digest of other people's links.

Unsubscribe anytime. We never share your address.

Next step

Have an idea? We will turn it into a system.

Thirty minutes, no slides. You will leave knowing what we would build first and what it costs to keep running. If we are not the right team for it, we will tell you on the call.

Or email us directly at hello@uxatom.com