Description
Summary ABOUT THE ROLE We are a small product team. Most of our code is now written by AI agents. Once writing code stops being the bottleneck, the hard problems move somewhere else: what is worth building at all, what an agent can safely decide on its own, and how you actually know the output works. That is the job. WHAT YOU WOULD OWN - Agents that operate as real teammates inside our product and our internal workflows, end to end. - The integration surface: APIs, webhooks, third-party platforms. - The permissions and approval layer that decides what an agent may do unattended versus what needs a human. - Evaluation and observability, so we can prove an agent actually did the job. YOU ARE PROBABLY A FIT IF - 6+ years building production backend or full-stack systems (TypeScript / Node). - You have shipped something real with LLMs or agents - not a demo. Something with users and failure modes. - You have worked on a platform, an SDK, or a developer-facing API before. - You are comfortable owning an area alone, without a manager handing you tickets. NICE TO HAVE - Experience with the monday.com platform, its apps framework, or marketplace apps. - MCP, tool-calling, or multi-agent orchestration. HOW WE WORK - Small team. Full ownership of your area. No sprint ceremonies, no platform on-call rotation. - Long-term engagement, and contract-to-hire is genuinely on the table. - Async by default, but I am responsive. TO APPLY - PLEASE READ Applications missing any of these will be skipped. Please skip the generic proposal. 1. REQUIRED: a short intro video. Loom, or any link that works. About two minutes, just you talking - no slides, no editing. I want to know I am hiring a human rather than a proposal generator. No video, no reply. 2. One agent or LLM system you built - specifically what broke in production, and how you found out. 3. Where you personally draw the line between what an agent decides and what a human has to approve. Start your reply with the word ORCHESTRA so I know you read this.