medoid.dev designs, builds and ships automation infrastructure, data pipelines and AI-assisted tooling — for enterprise, embedded, automotive and UAV domains where correctness is not optional.
We combine deep engineering expertise with pragmatic product thinking — so you get reliable systems, not impressive slide decks.
Web and end-to-end automation with Playwright/TypeScript, GUI testing with Squish, and protocol simulation with pymavlink — built to run in CI, not just on your laptop.
QEMU, Docker, Android AVD — reproducible environments that mirror your hardware so testing isn't a guessing game.
Azure DevOps, Jenkins, GitLab CI, Zuul — pipelines that give your team confidence to ship fast without breaking things.
CODEX and Claude integrated into the workflow to generate scaffolding, optimize test logic and cut boilerplate — not replace the engineer.
MAVLink, automotive IVI, embedded Linux — we've done the hard protocol work and know where the edge cases hide.
Onboarding, best practices, automation strategy and mentoring — so your team can own what we build together.
Real projects, real constraints. Names anonymised where required.
Our current engagement covers a system used by lawyers across the UK. We develop and maintain end-to-end test automation in Playwright and TypeScript, complement it with manual testing, and integrate the tests into Azure DevOps pipelines. CODEX and Claude support the day-to-day workflow for test creation, refactoring and maintenance, with engineering review and decisions kept firmly in the loop.
The client needed reliable, repeatable tests for a UAV ground control station — but hardware availability was a bottleneck. We built a pymavlink mock/simulator that models MAVLink communication deterministically, then layered Squish GUI tests over QGroundControl running inside a Docker-based virtual environment. Result: full regression coverage runnable in CI without a single drone in the room.
A premium OEM needed to validate an instrument cluster and HUD on Yocto-based embedded Linux before hardware was available. We designed Squish automation in Python, provisioned QEMU virtual environments mirroring the target board, wrote Yocto/Bitbake recipes to ship automation tooling alongside the product, and wired everything into Jenkins and GitLab CI pipelines with full artifact reporting.
Long-term engagement covering full CI/CD environment design (Jenkins, Gerrit, Docker, GitLab), Bitbake build system ownership, Robot Framework automation and performance testing, and cross-team integration coordination across multiple product lines. Delivered as Group Manager and Expert Engineer.
Tell us what you're building, where it's breaking, or what you want to automate. We'll tell you honestly what's worth doing, what isn't, and where to start.
[email protected]