James Walker, PhD
I take new ideas to revenue.
I build the first version of new products — the working build that closes your round, lands your first customers, and gets into real use. Deep on AI when it matters, plain software when it doesn't. PhD in machine learning, former CEO, bootstrapped operator — now building for founders and teams with the funding and the urgency to ship.
Who I am
An operator with research-grade depth.
I've spent the last decade in software — as an engineer, a CEO, and a researcher. I came up at an enterprise software business that was later acquired, then bootstrapped my own for seven years as CEO without venture capital, with customers across multiple continents. My PhD is in machine learning, with 500+ citations in the field.
I'm hands-on in production code most days — and equally comfortable in a board meeting or in front of investors. I've shipped research-grade systems to enterprises that pay real money. I know which ideas are worth building and which ones aren't.
What I do
Engagements I take on.
For founders and teams with funding and a deadline — just raised, about to, or launching a new product line. Each engagement is scoped in conversation, not sold off a shelf.
Turn a rough idea into a plan you can fund and build against.
A short, paid engagement that produces what you actually need to move: the system shape, the real cost and timeline, the shortest path to a working v1. Often the document that de-risks the raise — and the natural first step into the build.
Build the first version — fast.
Embedded with you from concept to shipped product. Founding-engineer depth, AI-heavy or plain software, whichever the problem actually wants. The working build your round, your first customers, or your launch is counting on.
Be your fractional CTO through the build.
The credible technical seat in the room — including in front of your investors — while you're shipping the first version and don't yet have a senior team. Monthly, scoped to what the stage needs.
Architect it so it survives contact with real users.
Senior judgement on the system shape — data, services, infrastructure, and where AI earns its place — before your team commits to code that's hard to back out of.
How I build
The stack I ship with.
Current practice, not aspiration. Everything here I've shipped to paying users.
AI & ML
- Anthropic & OpenAI APIs
- Claude Code / agent loops
- RAG architectures
- Vector databases
- Eval harnesses
Engineering
- TypeScript
- Python
- Next.js / React
- Postgres + Drizzle / Prisma
- Tailwind / shadcn
Production
- AWS / GCP
- Docker
- CI/CD
- Multi-tenant SaaS
- Monitoring & logging
What I've done
Selected work.
What I've actually shipped. Outcome first, stack second.
Where it runs
Systems I've built run in production at banks, telecoms, and global enterprises.
Got something to build, and a clock running?
Just raised, about to, or launching a new line — email me or grab 30 minutes. No pitch, just a fit conversation.