I'm Tyler Giles. I toured as a musician across 10 countries and most of the US, ran live A/V production on the road, then moved into product data, where I rebuilt my own job around Claude and Claude Code. I find the work that's eating a team's week and build AI tools that take it off their plate, with a human checking the results.
Email me LinkedIn
Things I've built that people rely on.
My job is keeping a large product catalog accurate. I mapped the manual process, then built a dashboard that runs several Claude Code agents side by side, tracks every supplier file from intake to import, and sends every result to a human-review queue before it goes live. A teammate uses it now too.
Roughly 60% more work done each week, on harder files.
Campaign staff told me what they needed. I designed and shipped a production web app with Claude Code: volunteer sign-up with personal links, an organizer dashboard, and door tracking that works offline on volunteers' phones.
9,300+ addresses grouped into 140 walkable door packets that never cross a major highway.
The startup I lead AI for started working with a newly launched model, and there was no reliable reference for it. I built a Claude skill and knowledge base from the official docs, shared across Claude Code and Codex.
Every session starts from verified facts instead of guesses.
Work tools are described, not shown. Company data stays at the company.
Start with the work, not the tool.
Watch how the job actually gets done, and find the part that eats the week.
A working version people can try beats a perfect plan.
Everything gets reviewed before it goes anywhere. Automation earns more trust as it proves itself.
Before-and-after numbers, plain-language docs, and training so the team owns it.
I use Claude to build and Codex as an independent reviewer, so a second model checks the work before I trust it.
The touring musician who automated his own job.
Before AI, I lived on the road. I toured through 10 countries and most of the US, playing pedal steel and lead guitar for country and bluegrass bands and performing as a singer-songwriter. Then I spent three years running live A/V production with a two-person crew: load the van, drive the gear, set up, run the show. Nothing teaches you to stay calm, solve problems fast, and explain technology in plain language like a show that can't stop.
That habit carried into product data, and then into building AI tools. I'm not a traditional programmer. I design the system, direct the build with Claude Code, test every change, and run it for real.
Music, baseball and A/V show up in pretty much everything I do. Now AI does too.
Let's talk.
About AI, a project, or a process that's eating your week.
tyler@tylergilescreative.com LinkedIn