Dave fromReadingReading → Bay Area
About
A hands-on engineering leader, four years into industrial automation after two decades of SaaS, ecommerce and developer tooling — now mostly interested in what AI actually changes about the work.
Software roots. Industrial automation now. AI across both.
I have spent most of my career shipping web software; the last four years have been at Mytra, where the software tells material-handling systems what to do. Throughout, an unfashionable interest in how much verification a thing needs before you let it near anything expensive.
The through-line is proportion. A script that generates a PDF does not need the same rigour as the controller that moves a tonne of steel, and treating them identically wastes everyone’s time in both directions. I wrote about that here.
These days I lead an engineering organisation hands-on: writing and reviewing code, and working out where AI tools genuinely improve the way we build, review and ship. Being the person who has to answer for the system when it misbehaves at three in the morning tends to sharpen that judgement.
The path, roughly
- Industrial automation
- The last four years, at Mytra: software-defined material handling and storage for warehouse operations. I joined as one of the first engineering hires, built the server and UI foundations, and have led the organisation through the deployments that followed. Physical systems make the feedback loop obvious — a wrong assumption shows up in a facility, not a dashboard.
- AI in engineering
- The current argument: AI is cheap enough to produce code, which moves the question from whether a team can build something to whether it can tell what it has built. Most of my writing is about that gap — where AI earns its place in a codebase, and how much rigour the result has actually earned.
- SaaS, ecommerce & developer tooling
- Before that: workflow automation at Slack, developer experience tools at Walmart Labs, and web software at Rivian and Zappos. A long stretch of systems where the failures are quiet — a broken checkout, drifted data, a slow leak rather than a loud stop. That taught me release discipline, review practice, and a professional suspicion of anything described as provisional.
- US 9983855
- Systems, method, and non-transitory computer-readable storage media for displaying software components. Issued 29 May 2018 — filed years before anyone was arguing about whether AI could write software, which I find quietly funny.
- Loughborough, 2002–2005
- BSc (Hons) Information Management and Computing. Where the engineering degree and the tolerance for detail came from.
Working together
I am not looking for a full-time role. I am most useful as an advisor or a sounding board on problems where the hardware and the software are not separable — material flow, robotics orchestration, systems where the feedback loop is a physical one. I am also glad to talk through AI adoption in engineering: what actually changes, what doesn’t, and how to keep review and release discipline intact while the tools move under you.
The full professional record is on LinkedIn ↗. If any of that sounds like a problem you are working on, get in touch and tell me what you are deciding.