1. Close incoming platform tickets from Shared Services. Agents, Team Leads, and Operations report issues daily; quick fixes, workflow bugs, configuration problems. You triage, diagnose, and ship.
2. Find the pattern and fix it at the source. When multiple agents report the same friction, that's not a ticket, it's a product problem. When a client complaint traces back to repeated agent failures, you ask why the platform is allowing it, and you close that gap for good.

• Own features end to end; problem framing, design, backend, frontend, instrumentation, rollout, and the follow-up fixes nobody filed a ticket for.
• Work with founders, product, and design at the whiteboard stage, before anything is written down, and help decide what's worth building.
• Build the AI surfaces of the product; prompt and agent architecture, tool calling, retrieval, guardrails, and the human-in-the-loop escape hatches that make automation safe to ship.
• Build and maintain evals. Treat "does this actually work" as an engineering artifact with a regression suite, not a vibe check before launch.
• Watch real usage. Session replays, logs, transcripts, support tickets, and customer calls. Bring back what you learn as shipped changes, not as a document.
• Own the details that make software feel good; latency budgets, token cost, empty states, error copy, loading behavior, keyboard paths. These are part of the feature, not a separate polish phase.
• Cut scope and kill things. Argue a feature is wrong before you build it well.
• Compound the team's velocity; CI, preview environments, internal tools, seed data, debugging surfaces. Leverage counts as product work.

• Not a ticket queue firefighter. That's other roles on the Operations side.
• Not a shadow engineering team. You report to and work within Engineering.
• Not an independent operator building outside the platform. Everything you ship goes through product.

• You want a ticket fully spec'd before you'll touch it
• You treat "why does this keep happening" as someone else's question to answer
• You measure your work by whether you closed the ticket, not whether the pattern stopped
• Your AI/LLM experience is prompting a demo — never held to a reliability bar with a real agent depending on it at 2am

• 4+ years of full stack (stronger on the BE) software engineering experience shipping production enterprise software
• Direct experience building internal tools, agent-facing platforms, or operational systems
• Product mindset: comfortable scoping problems, asking the right questions, and communicating/prioritizing tradeoffs
• Track record of shipping fast while working inside a product and design process
• Experience working asynchronously with distributed teams

• Strong fluency in TypeScript. We currently use:
  • Next JS
  • NestJS
  • TypeORM with Postgres
  • BullMQ
• Comfortable with agentic coding tools (Claude Code, Cursor, Replit Agent)
• API integration experience (REST, webhooks, third-party SaaS APIs)
• Prompt engineering experience with production AI systems (OpenAI, Anthropic, or similar)

• Experience building tools for contact center, customer support, or operations teams
• Short-term rental, hospitality, or property management domain knowledge
• Track record of deploying AI agents in production

• Reports to the Head of Engineering
• Priorities set jointly with the Head of Operations, based on operational signal
• Product decisions filter through the Head of Product
• Design decisions go through our design team
• Weekly syncs with Engineering, Operations, and Product
• 5 hours of overlap with Miami time (Eastern), minimum, daily

Compensation

Base range is $70-105K USD depending on location and experience.