Work that shows what the team can build
How we approach hard engineering problems, from peer-to-peer networking to mobile apps, infrastructure and the security that runs through all of it.
Engagements are kept confidential. The work below is shown as representative projects, anonymised by industry, with the Vexaro platform itself as our flagship build.
Representative projects
Each is a real engineering problem and what we built to solve it. We frame outcomes by capability and architecture, not invented numbers.
Vexaro: remote access over peer-to-peer
- Challenge
- Connect two machines directly across hostile NATs without routing every session through a datacentre, while keeping latency low enough for live screen control.
- What we built
- A Rust and Go platform that connects directly wherever the network allows and through an authenticated relay when it doesn't, with every session encrypted and identity-verified.
Remote sessions responsive enough for live screen control, direct where the network allows and relayed where it doesn't. This platform runs on that architecture.
Cross-platform client for a field-operations team
- Challenge
- One codebase for iOS and Android that stays responsive on weak networks and survives being backgrounded mid-task.
- What we built
- A Flutter app with typed state, offline-tolerant sync against a Go API, and a release pipeline that ships both stores from a single build.
A single team owns both platforms; new features land once and reach every device, with no native fork to maintain.
Identity, access, and audit for an organisation tier
- Challenge
- Let organisations bring their own identity provider, scope every action to a tenant, and answer 'who did what, when' for compliance.
- What we built
- Standards-based single sign-on (SAML, OIDC) and directory sync (SCIM, LDAP, Entra ID), tenant-scoped authorisation across the API, and an audit log with CSV export and SIEM forwarding.
Access maps to the customer's own directory, and changes made through the API are recorded in the audit log with who made them and when. That is the foundation for serving teams as well as individuals.
Platform foundation that scales without surprises
- Challenge
- Run a stateful product reliably, deploy many times a day, and see what's happening when something breaks.
- What we built
- For the client we built containerised services on Kubernetes, described as code, with managed Postgres and Redis, blue-green releases, and metrics, logs and traces wired in from day one.
Environments are reproducible and recoverable; the same definition stands up staging and production with no manual drift.
Security built into the pipeline, not bolted on
- Challenge
- Catch vulnerable code and dependencies before release without slowing the team to a crawl.
- What we built
- Static analysis and dependency scanning on every change, hardened build pipelines, and a penetration-test pass before each major release.
Whole classes of issue are caught at merge time; security is a property of the process rather than a one-off review.
Adaptive video for live remote control
- Challenge
- Keep a remote desktop sharp and smooth as bandwidth and round-trip time swing, without a heavyweight media stack.
- What we built
- A Rust real-time video path that adapts to the network.
The stream adapts quality and frame rate to the link in real time.
Capabilities
The languages, frameworks, and infrastructure we work in every day.
Languages
Frameworks & runtimes
Infrastructure & data
How we work with you
Project build
A defined scope, a fixed team and clear phases, from discovery to launch and support. Best when you know what you need built.
Embedded team
Our engineers work inside your process and tooling, extending your team's capacity without the overhead of hiring.
Audit & advisory
A focused review of an architecture, codebase, or security posture, with a written, prioritised set of recommendations.