We're not actively hiring.
Eastsoft is a small, deliberate team. We grow by inches, not headcount targets — and when we do open a role, it's usually filled by someone we already know. But if you build at the protocol layer of AI, we want to know you exist.
How we actually hire.
Eighteen years in, Eastsoft is still a tight team. We don't run job boards, we don't post evergreen reqs on LinkedIn, and we've never used a recruiting agency. When a role opens, it's because the work demanded it — and we usually already know who fits.
That said, we always read open applications. If you've shipped something at the protocol layer — a parser, a small language, a compiler, an agent runtime, a paper that an LLM team actually cited — write to us. We answer.
What we look for.
Taste at the boundary
Where syntax meets semantics. Where humans meet machines. Where a thing could be three lines or three hundred. You pick three.
Care for the spec
You read RFCs for fun. You've been bitten by ambiguity in a standard and went back to fix the standard, not your code.
Quiet, finished work
You ship. You don't ship loud. You measure twice. The PR description is shorter than the diff because the diff explains itself.
Originality
Eastsoft work tends to look slightly weird the first time you see it. If your portfolio is "the same thing every other startup makes" — we are probably not the right fit.
Tech & practice.
Working at Eastsoft.
Remote-first, Canadian payroll, asynchronous by default. Generous compute budgets, generous reading budgets, generous quiet. Few meetings. No "all-hands theatre." Compensation is competitive but the headline benefit is the work itself: shipping standards-grade tooling that real model providers cite.
We will never ask you to "move fast and break things." We will absolutely ask you to move carefully and finish things.