Portfolio
Every tool below is designed, built, and run by me, in-house - real production systems, backed by real automated tests, not tech demos.
Get in touch via the contact form and I'll talk you through it.
Fuel Strategy App
Live fuel strategy for iRacing team endurance events - splits remaining fuel across stints and tells the driver what to add at the next stop, in real time.
ATLAS HWiNFO Bridge
Streams live hardware sensor data - CPU, GPU, VRMs, fans, drives - from HWiNFO into Motion Applied ATLAS for telemetry-style analysis of the rig itself.
ATLAS System Stats Bridge
The same idea for CPU, RAM, disk, and network - so overall PC health sits alongside hardware sensors in the same ATLAS trace.
ATLAS Monitoring
A desktop watchdog for the ATLAS stack itself - keeps Docker, Kafka and the Stream API healthy and restarts what falls over, so a race session doesn't lose telemetry.
Office Lights
A local desktop app for Philips Hue and Tapo cameras around the rig - no cloud account, everything stays on the LAN.
ATLAS Control - Stream Deck plugin
Drives Motion Applied ATLAS's keyboard shortcuts from an Elgato Stream Deck - 158 commands catalogued, with focus-safety and hold-to-repeat built in.
ATLAS Constants Switcher
Swaps a running ATLAS 10 session onto a different car or track's Function constants in seconds - no restart, no disrupting engineers already working in ATLAS.
racestrat.uk
This site's sibling - a team endurance racing site with live driver stats, media, and the Fuel Strategy App write-up, on Cloudflare Workers, D1 and R2.
What ties these together
Every one of these exists because something about running a sim rig and a race weekend was annoying enough to fix properly: no way to see rig health next to lap data, no way to trust the ATLAS stack was still running mid-session, no way to dim the rig lights without reaching for a phone app, no way to drive ATLAS from a Stream Deck instead of a keyboard reach mid-corner. Same pattern each time - identify the actual friction, build the smallest thing that removes it, and back it with real tests rather than "it worked once on my machine."