SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Luna is a second-generation autonomous agent platform built inside monday.com's Agent Labs, a program that spins up independent startups from scratch. Luna operates as a seed-stage company with a tiny team, real product, and real users backed by monday.com's resources and distribution.
Luna has already shipped a live product and is now solving the hard problems that first-generation agents exposed: unbounded self-coding, execution drift, and lack of observability. The second-generation Luna runs agents on deterministic workflows with built-in trust mechanisms: agents write and refine workflows that run on a real engine rather than re-reasoning each time, a one-way vault for secrets that never enter the model's context, rollback and audit trails on every action, and approval gates for sensitive operations. Everything else—tools, channels, skills—is pluggable and isolated per agent.
You will be the engineer who builds this product. You work directly with the founder (monday.com's CEO), with no layers between you and decision-making. Decisions happen in conversation, not in roadmap reviews. You own the engineering; the founder owns the product. There is no playbook yet for what a good agent is in this context—you'll be writing it with real use cases and real constraints.
Luna is not a chatbot. Users describe a goal, Luna builds the workflow, users correct it, and it runs reliably next time. A private Luna instance spins up in about a minute today.
REQUIREMENTS:
- Experienced builder at heart; you prioritize shipping over pitching. This is an engineering role, not a founder-track role.
- Strong instincts about agents: before writing code, you think about what the model must never touch, how you'd know it failed, and what happens when it does. Shipping ten agents is not required; having thought hard about one is.
- Independent operator: given something vague, you return with something that works.
- Strong fundamentals in distributed systems and reliability, with production experience.
- Nice-to-have: LLM tooling in production, tool use, sandboxing, evals, and a security-minded approach to building.
