Enterprise AI is the biggest infrastructure problem of the current era. We're building the governance layer from the ground up — and we want engineers who want to own a piece of that.
See open rolesCulvii Kit requires cryptography (DIDs, Ed25519, Merkle Trees), distributed systems (immutable audit infrastructure, multi-tenant isolation), and language-level API design (TypeScript + Python SDK primitives that have to be right from day one). There's no shortage of genuinely hard problems.
We're a small team. Everyone owns something real. If you join us to build the HITL system, you own the HITL system — architecture, implementation, docs, and everything that breaks in production. No hand-offs to a PM who then hands off to a tech lead who then assigns a ticket.
Our engineering blog is public. The changelog has author attribution. The SDK has contributors. If you build something excellent, it's visible — to the community, to future employers, to anyone who reads the code.
We work directly with engineering teams at JLL, Visot, and similar companies. You'll understand the regulatory and operational constraints of real enterprise deployments, not a hypothetical spec. That context makes you a better engineer.
Early equity. We're early. Equity means something here. We won't obscure the details.
We have opinions about API design. Strongly typed primitives. Consistent patterns across both languages. Documentation is part of the code review, not an afterthought.
We use PR templates. We write descriptions that explain the why, not just the what. Code review is a collaboration, not a gate.
docker-compose up and you're running. No tribal knowledge. If setup takes more than five minutes, we fix it.
We ship every sprint and write about it. The changelog is not a release note — it's a demonstration of velocity. Author attribution is required.
We don't confuse presence with productivity. Async communication by default. Synchronous time is reserved for things that genuinely need it.
You won't wait 6 months to ship something meaningful. The entire SDK is a few people. Your fingerprints are on the product from week one.
The TypeScript SDK is the primary surface through which enterprise engineers interact with Culvii Kit. You'll own the Actor, Workflow, Tool, and HITL primitive implementations — their API design, their correctness, and their performance characteristics.
Bonus: experience with cryptographic primitives, DID standards, or compliance-sensitive systems.
The identity and registry layer — DID issuance, Ed25519 key management, capability registry, trust elevation flows, and the Merkle Tree-based audit system. The most cryptography-intensive part of the codebase.
Full Python SDK parity with the TypeScript SDK. The Python SDK is critical for data science and ML engineering teams who build agents in Python-first environments (LangChain, Jupyter, LlamaIndex).
Benchmarked to senior engineers at top-tier companies. We pay well.
Early stage, real equity. You're an owner, not an employee.
Work from wherever you do your best thinking. We're async-first.
Full coverage. Details on request — we're not hiding anything.
Set up a workspace that doesn't make you miserable.
Conferences, courses, books. If you'd present or learn something relevant, we'll fund it.
Total time investment: roughly one week, mostly async.
Short email exchange or async video. We want to understand what you're working on and what you care about. Not a screening quiz.
A real technical discussion about a problem relevant to our stack. Not leetcode. A conversation between engineers.
A small, real, paid project. You'll solve something we've actually encountered. Scoped carefully. You keep the payment regardless of outcome.
Meet the team, ask hard questions, decide if you want to work here. We'll be honest about the stage and what it's like.
We don't care about your college degree, your current employer's logo, or your LinkedIn follower count. We care about what you've built, how you think, and whether you want to work on this problem.
Include: what you're working on now · something you've built that you're proud of · why this problem is interesting to you