James Jewhurst — Engineer / Architect / Builder

I build software that makes complicated things feel simple.

I'm James. I build software for real people and real workflows, from automotive platforms and integrations to native apps and developer tools. I care about useful products, sound architecture, and the unglamorous engineering work that makes software hold up in the real world.

Now: Principal Engineer, Shift Automotive
Before: Director of Engineering, VolieSVP of Technology Operations, Bolt On Technology
Public: 6 projects

See the work Contact

A few things I've built

These are the decisions.

AI Usage MeterTue 1:03 PM

01 · macOS · iPhone · WatchShipping

AI Meter

See Cursor, Claude Code, and Codex usage without living in three dashboards. A local-first meter for included pools, rolling windows, spend, and on-demand.

Status: Signed and notarized Mac builds on GitHub Releases. iPhone and Watch stay on TestFlight.
Choice: Credentials stay in the device Keychain. The meter reads local sessions. It does not send them to a backend I run, store tokens in iCloud, or hand credentials to the Watch.

Case study Source aiusagemeter.app

02 · macOS 14+ · Apple siliconShipping

CursorStack

All those Cursor windows, one manageable workspace. A tab strip over the real windows, rather than another editor.

Status: Signed and notarized Mac builds on GitHub Releases. Apple silicon, macOS 14 or newer.
Choice: macOS Accessibility moves, resizes, and focuses the original Cursor windows. CursorStack does not embed Cursor, read source, or upload anything.

Case study Source cursorstack.app

03 · iPhone · Watch · MacPublic source

Daily On Plan

A daily nutrition sheet for iPhone, Apple Watch, and Mac. Protein toward a goal, hydration, and whether you followed the plan. Not a service, and not a medical device.

Status: Public source. The README does not establish a public App Store release.
Choice: On the device by default. Optional iCloud sync for the journal. No account, and no backend of mine. iCloud is still a cloud feature.

Case study Source

DoteqPython · MIT
A Python tool that adds missing keys from an example env file without overwriting the values you already have.
GLauncherWindows
A Windows app that opens a group of applications together, in the order you set, instead of starting each one by hand.
BudMathWeb
A small web calculator for ounce and gram combinations, with history kept in the browser.

All work On the bench: DevPulse, HookBench, PocketPTY

Practice

Software for the people who keep the real world moving.

Much of my professional work has been in automotive software, on both sides of the industry: dealerships selling and servicing new cars, and the independent repair shops of the aftermarket. Repair shops are complicated places. The work crosses scheduling, estimates, parts, technicians, customers, and systems that do not always speak to each other. That shaped how I think about products. Respect the workflow, integrate carefully, keep things understandable, and make the software reliable when the day gets busy.

Product and architecture
Connect the decision to the system that has to live with it.
APIs and integrations
Make systems that do not share a vocabulary still hand work off cleanly.
Engineering leadership and delivery
Ownership, process, and the path to production are part of the software.

Software should earn its complexity

Solve the real problem.
A good implementation of the wrong idea is still the wrong idea. Understand the person doing the work and what would actually help.
Start small, and design to last.
Ship a coherent first version and leave room for what real usage teaches you. Do not prematurely build the platform for version ten.
Care about the details that matter.
Performance, reliability, privacy, recovery, accessibility, and deployment are part of the product, not polish added after the demo.
Own the whole thing.
Product understanding, architecture, implementation, operations, and the user experience are connected. Taking responsibility across those boundaries produces stronger software.
Prefer clarity over cleverness.
If the straightforward version works, that is the version. Cleverness has to earn its place the same way complexity does.
Use AI deliberately.
AI can compress drudgery, speed research, and support coding. Judgment, verification, security, and responsibility still belong to engineers.
Keep learning by making.
Small experiments and side projects sharpen the thinking that goes back into larger systems.
The longer version

I build software that makes complicated things feel simple.

A few things I've built

AI Meter

CursorStack

Daily On Plan

Practice

Software for the people who keep the real world moving.

Product and architecture

APIs and integrations

Engineering leadership and delivery

Software should earn its complexity

Solve the real problem.

Start small, and design to last.

Care about the details that matter.

Own the whole thing.

Prefer clarity over cleverness.

Use AI deliberately.

Keep learning by making.

I still get excited about a good problem.