About the job

You will build AI features end to end: the model or pipeline that produces the decision, the service that runs it, and the product surface where a revenue manager sees and acts on it. This is not a research role and it is not a pure app-dev role — it is the bridge between them.

Examples of what you'd own in your first year

• Forecasting and rate-recommendation logic that feeds our auto-publish engine, including the guardrails that keep recommendations inside a property's tolerance
• LLM-driven features: natural-language explanations of why a rate moved, conversational querying of a property's performance, automated market commentary
• Rate-intelligence pipelines — ingesting competitor rates, cleaning messy third-party data, and turning it into signals the model can trust
• Evaluation harnesses and monitoring so we know when a model degrades before a customer does
• The full-stack surface around all of the above: APIs, data models, and the UI that exposes them

Requirements

• 3–5 years building and shipping production software, with meaningful AI/ML work in that time
• Strong Python; comfortable in TypeScript/JavaScript for the application layer
• Hands-on experience with LLM APIs (Anthropic, OpenAI or similar) — prompt design, structured output, tool use, retrieval, and the evaluation discipline to know whether any of it is working
• Solid SQL and data modeling; you can reason about a schema and write queries that don't melt the database
• Experience with at least one of: time-series forecasting, pricing/optimization models, or recommendation systems
• Comfort with messy, incomplete, real-world data — third-party feeds, inconsistent APIs, partial history
• Cloud deployment experience and familiarity with modern CI/CD

Nice to have

• Hospitality, travel, pricing, or marketplace domain experience
• Experience integrating with PMS, channel manager, or booking-engine APIs
• Background in a small startup where you owned a system rather than a ticket queue

How we work, and who succeeds here

We hire for ownership as much as for stack. The engineers who do well:

• Take technical ownership. You're handed a problem, not a spec. You decide what to build, flag the trade-offs, and are accountable for the result in production — including when it breaks at an inconvenient hour.
• Are intrinsically driven. Nobody will watch your calendar. You choose the right next thing to work on when the direction is ambiguous, and you finish things.
• Communicate well asynchronously. Written updates that a busy reader can act on. You surface blockers early rather than going quiet. Clear, complete PR descriptions and issue comments.
• Manage your own time zone. You're responsible for maintaining reliable overlap with the team and for being predictable about when you're reachable.