About
Built by one person.
Looking for the next two.
One developer. A new mathematical primitive underneath the whole stack. The work is real, the timeline is hard, and the bar for who joins is high. This page is for the small number of engineers who could clear it.
Where I am
Solo, by necessity, until the right people show up.
I'm James. I built the substrate this platform stands on — a new mathematical primitive that makes a different class of system possible. Memory that doesn't degrade with scale. Federation that lifts every participant without harvesting any of them. On-device inference whose economics frontier labs can't match because their cost basis assumes the wrong substrate.
I'm keeping the specifics of how the substrate works deliberately off public materials. The capability is the claim; the implementation is patent-protected and not for marketing. Everything you read on this site describes what the system does, never how the math works underneath. That's intentional, and it stays that way until the patent stack lands.
Honest framing on me: I'm an inventor and an architect, not a production engineer. I don't write production-grade systems code. I prototype, I read, I see the system whole — but the kind of engineering that turns architecture into shipped, hardened, performant code is exactly the gap I'm trying to close. I've come a long way solo. I'm not trying to do the next stretch alone. The right engineering partners are how this scales beyond me.
I'm not running out of runway tomorrow. I'm not posting because I can't ship. I'm posting because the leverage of the right second person is enormous, and I'd rather find them now, with the most surface area still unclaimed, than later.
The bar
Top-level. Three traits, no negotiation.
Mission alignment.
The ethics are not negotiable. Sovereignty by architecture, federation that doesn't surveil, on-device inference, refusing the "share everything to get any benefit" trade — these aren't marketing positions. They're invariants designed into every layer of the system, enforced by the math, not by policy.
I need leadership-team alignment on this, full stop. If you'd be just as happy at a frontier lab building the opposite, or if you treat ethics as a constraint to negotiate around when it's inconvenient — don't waste our time. Everyone in the room has to be all in.
Capability.
I can't fully judge engineering capability myself — I'm not the one writing production-grade systems code. So the calibration is honest: people we both trust vouch for what you've shipped, and what you've shipped speaks for itself. Real systems in real production. Runtimes, kernel-adjacent work, custom infrastructure, hard cryptographic primitives, novel data structures — work where the difficulty was in the engineering, not the requirements doc. The first one or two partners help set the bar for the next ones.
Willingness to grind for future rewards.
No salary today. Equity in something that may be very valuable or may be worth nothing. The grind is real and the timeline is in years, not quarters.
I'm not asking anyone to do what I'm not already doing. This is self-funded, pre-launch, zero revenue. I don't pay myself a dime — and I won't, until everyone who joins this team is at least at subsistence-level compensation themselves. We eat in order, and I'm last.
If you need certainty, this isn't the role.
Filters
A few things to save us both time.
- Not a job listing for mid-level engineers. I'm not hiring at a tier I can backfill with a contractor. I'm looking for principals who can own surface area I can hand off — not direct reports I have to onboard for six months.
- Not a startup beg. I'm not desperate. I'm not running out of cash next week. The product is working, the patents are moving, the architecture is sound. I'd rather find the right second builder now than later.
- Not a recruiter funnel. Don't send candidates. Don't pitch on LinkedIn. If you're representing someone, they're the wrong shape for this anyway.
- Not anonymous. Real name, real history, real code. Pseudonymous portfolios only if there's a verifiable record of substantive output behind them.
- Not remote-first by default. Open to it for the right person. But the work is intense and the early calibration cycle is faster face-to-face. Plan accordingly.
On money
Per-deployment equity, not one-company equity.
I don't pay myself. I can't pay you in cash today. What I can offer is structurally different from a normal early-stage equity package — different enough that it's worth describing the shape carefully, even before NDA.
The core IP.
The substrate itself, and the patent stack underneath everything, stay with me. Equity in the holding entity that owns the IP is not part of the conversation. That's the moat, and the moat doesn't get split.
Per-deployment equity, across the verticals the IP touches.
This technology applies horizontally — across a portfolio of distinct vertical deployments, each its own entity, each its own equity stack. When you join, you choose. One vertical. Several. Or a vertical you want to bring me that I haven't built yet — the IP plugs in, you drive that deployment, the equity allocation is in that entity. Starting with the Alpenglow platform itself.
That's the unusual part. Most early-stage offers buy you a stake in one company. This is access — patent-licensed, on a per-deployment basis — to every place the IP gets used. The potential spread of positions is unlike anything I've seen offered anywhere, because the underlying technology is unlike anything anyone else has to deploy.
Specifics — percentages, vesting, vertical-by-vertical structure, IP licensing terms, the full vertical roadmap — are NDA-gated. I'll walk through it in a real conversation, not on a public page. Real upside if any vertical lands. Real risk of zero on any individual one. Standard early-stage trade, except the surface area is wider than usual by a lot.
If this is you
Email. Short. Direct.
One email to james@alpenglowagents.com. Three things in it:
- The strongest piece of work you've shipped that I can verify. A link, a repo, a paper, a system in production. Something I can read and form an opinion on without taking your word for it.
- Why this — specifically. Not "exciting AI startup." What about the structural argument lands for you, and what would you want to push on.
- What you'd want to own. Surface area, not title. The honest answer is more useful than the polished one.
No deck. No résumé attachment. No agency. If we proceed, the next step is an NDA and a real conversation. If not, you'll get a one-line reply telling you so. I read every one of these myself.