Skip to main content

Engineering digital infrastructure for organisations that matter.

We design, integrate, build and operate the platforms that governments, development programmes and enterprises depend on -- and we stay accountable for how they run afterwards.

Financial management platforms

Budgeting, approval workflows, reporting and loan tracking, replacing spreadsheet-driven processes.

Fleet operations

Vehicle, trip, fuel and maintenance management with fleet analytics and a mobile app for field workflows.

Government and health systems

Among the business systems delivered since 2022, alongside insurance broker platforms.

Engineering behind the scenes

Architecture, integration and platform engineering delivered in partnership with other technology providers.

Based in Lilongwe, Malawi

Working with clients in Malawi and internationally.

Selected work

All work
Confidential enterprise clientEnterprise

Enterprise Financial Management System

Replacing spreadsheet-driven finance with one system: budgeting, enforced approval workflows, reporting, document management, and loan tracking with automatic interest calculation.

Problem
The organisation ran its finances on manual processes and spreadsheets. Records were fragmented, reporting was late, audit trails were weak, the same data was entered more than once, approvals stalled, and there was a standing risk of losing files outright. Nothing enforced an approval sequence, and finding a specific piece of information took longer than it should have.
Outcome
Fewer manual errors
  • digital platforms
  • integration
Confidential enterprise clientEnterprise

Fleet Management System

Vehicle, trip, driver, fuel and maintenance management with GPS and telematics integration, a mobile app for field workflows, and analytics built for fleet cost and reliability decisions.

Problem
Vehicle booking and allocation were manual, availability was hard to see, maintenance tracking was weak and fuel records were missing or incomplete. Trip approvals were inefficient, drivers and vehicle assignments were hard to follow, and reporting could not answer the questions that actually drive fleet decisions: which models fail most, which cost most to maintain, which vendors are reliable.
Outcome
Visibility of the fleet
  • digital platforms
  • integration
  • data ai

Why Cosmic Limit

What actually differs

  • Architecture before code

    We establish the data model, the boundaries and the failure modes before building, because those are the decisions that are expensive to reverse.

    We define the solution architecture, system boundaries, integration approach and deployment model before implementation begins -- as in both published engagements.

  • Integration is the default assumption

    Almost nothing we build stands alone. We design for the systems already in place, including the ones that cannot be changed.

    We design for external integrations from the outset: third-party platforms, GPS and telematics providers, and client-controlled systems with limited availability or restricted access.

  • We operate what we build

    Deployment, observability and recovery are part of the engagement, not a handover problem. A system that nobody can run is not delivered.

    We stay involved beyond implementation through ongoing system maintenance, application support and operational improvements -- including for both published engagements.

  • Built for the conditions it will run in

    Intermittent connectivity, constrained infrastructure and small operating teams are design inputs rather than edge cases.

    We design for the operational realities of the target environment: poor connectivity, strict security restrictions, client-controlled infrastructure and unreliable external systems.

How we work

Discover, architect, build, validate, operate

  1. 01

    Discover

    Understand the process as actually performed, the constraints that bind it, and what success would look like to the people accountable for it.

  2. 02

    Architect

    Establish the data model, service boundaries, integration points and security model, with the trade-offs written down.

  3. 03

    Build

    Deliver in thin slices that reach production early, so integration and operational problems surface while they are cheap.

  4. 04

    Validate

    Test against the real conditions -- data volume, connectivity, accessibility and the people who will use it.

  5. 05

    Operate & improve

    Deploy, observe, respond and hand over, with the documentation and runbooks an internal team needs to take it on.

Engineering credibility

How we build, concretely

Principles

  • Boring technology, deliberately

    We choose the least novel option that meets the requirement, so the organisation can hire for it, support it and outlive us on it.

  • The build is reproducible

    Containerised, pipeline-driven and portable, so hosting stays a decision rather than a lock-in.

  • Decisions are recorded with their reasoning

    A future team can tell whether a choice still holds or was made against conditions that have since changed.

  • Accessibility is a build requirement

    WCAG 2.2 AA is part of the definition of done, not a remediation phase.

Practices

  • security

    Server-side validation, least-privilege credentials, dependency and licence scanning in the pipeline, and no secrets in the repository.

  • performance

    Performance budgets enforced in CI rather than measured after launch, because a regression found in production is already someone's problem.

  • reliability

    Health and readiness separated, recovery procedures tested on a schedule, and alerting tied to conditions that matter.

  • accessibility

    Keyboard-operable throughout, contrast verified in both themes by automated test, and reduced-motion respected.

Stack

Application

  • TypeScript
  • React
  • Next.js
  • React Native
  • Node.js
  • Python
  • .NET

Data

  • PostgreSQL
  • Redis
  • RabbitMQ
  • dbt

Cloud

  • AWS
  • Azure
  • Google Cloud

Platform

  • Docker
  • Kubernetes
  • Terraform
  • Linux
  • Nginx

Practice

  • Playwright
  • Vitest
  • GitHub Actions
  • OpenTelemetry

Who we help

Where this work lands

  • Government & public sector

    Government & public sector

    Registries, service delivery platforms and reporting systems, built for interoperability, auditability and long-term maintainability by the institution that owns them.

  • Development organisations

    Development programmes

    Monitoring, evaluation and reporting systems that stand up to donor scrutiny and keep working in the field conditions they are deployed into.

  • Enterprise

    Enterprise

    Integration layers, operational platforms and modernisation work that reduces technical risk rather than relocating it.

  • Growth companies

    Growth companies

    Product engineering for teams that need to get from a working idea to something that survives real usage.

Evidence

How we work, in practice

  • Every architecture decision is written down

    Including the options not taken and why, so the reasoning survives the people who made it.

  • You get the constraints, not just the recommendation

    We tell you what a choice forecloses and what it would cost to reverse.

  • Handover is designed in from the start

    Runbooks, environment documentation and the walkthroughs an internal team needs to operate the system without us.

Have a system, an integration or a data problem worth getting right?

Tell us the constraint you are working against. We will tell you how we would approach it, what we would need to know first, and whether we are the right people for it.