Staff Engineer, AI Automation

AvantStay

15h•Remote

About The Position

AvantStay delivers world class, authentic, tech-enabled short-term rental (“STR”) group experiences targeted at the millennial generation. We are venture funded and growing rapidly in the explosive $100+ billion dollar STR industry. We deliver a customized end-to-end experience that is tailored just for groups and powered with technology at every layer. We are seeking a Staff Engineer who will own AI automation across our business - building autonomous systems that monitor data, take action, involve humans when needed, and learn from feedback. We need someone who moves fluently between business and technology. On one side, that means going deep on how work actually gets done, spotting where manual effort burns time, introduces error, or simply doesn't scale. On the other hand, it means knowing the technical landscape well enough to translate those opportunities into systems that run in production: reliable, observable, and built to handle the messy edge cases that only show up in the real world. Some workflows are big (automating an entire listings lifecycle), some are small (a bot that flags pricing anomalies). We need someone who ships both. When you join, you'll work with a small, senior engineering team that has already shipped multiple AI agents to production - including lead qualification, data querying, revenue automation, and infrastructure monitoring systems. Our stack includes Scala and TypeScript across the backend, PostgreSQL, ClickHouse, Kafka, Kubernetes on AWS, and deep integrations with Claude, OpenAI, Pi, LangGraph, MCP and other tools for AI system management and integration. You won't be starting from scratch - you'll be accelerating what's already working. As a Staff Engineer, you'll report directly to engineering leadership, working alongside a team of ~30 engineers across backend, frontend, data, and DevSecOps. Our team is fully remote, flat, and talent-dense. We move fast and expect you to own outcomes, not just write code.

Requirements

Deep understanding of agentic workflows and the current tooling landscape - orchestration, evals, agent SDKs (Claude Agent SDK, OpenAI Agents API, Vercel, PI, LangGraph), tool-use patterns, multi-step reasoning, structured outputs, and RAGs.
Experience shipping production AI/LLM agent systems - not prototypes, not chatbots. Autonomous agents that take real actions in real systems.
Strong software engineering fundamentals - you can navigate databases, event systems (Kafka), API protocols (GraphQL, REST), containers (Kubernetes), and cloud infrastructure (AWS). Our systems are complex, and you need to be able to work through that complexity.
Our stack includes TypeScript and Scala (functional, Typelevel), Python - comfort reading and contributing to typed functional code is a plus, but deep Scala expertise is not required.
Product instincts - you can talk to a revenue manager or sales, understand their workflow, and figure out what to automate without being handed a spec.
Bias toward shipping. We value working software over documents.
Fluency in English, both written and spoken, for effective communication in a distributed team.

Nice To Haves

Background in property management, hospitality, or marketplace platforms.
Experience with functional programming (Scala, Haskell, Clojure, Elixir).
Experience with distributed systems, Clickhouse and Postgres.

Responsibilities

Design and build autonomous agentic loops across business domains - monitoring listings performance, qualifying leads, resolving customer tickets, optimizing SEO content, and many more we haven't built yet.
Own the full lifecycle: understand the business workflow, architect the agent, ship to production, monitor performance, iterate based on feedback.
Connect agents to our data infrastructure - PostgreSQL, ClickHouse, Kafka event streams, GraphQL APIs, MCPs, third-party integrations - so they can observe, reason, and act.
Build human-in-the-loop escalation patterns where full autonomy isn't appropriate.
Work directly with business stakeholders (revenue management, marketing, CX, sales) to identify automation opportunities and translate them into agent architectures.
Untangle complexity in our existing systems to make them agent-accessible - wrapping services, building APIs for agents, MCP servers, creating clean data interfaces, and transforming existing ones.
Establish patterns and tooling that help the broader engineering team deploy agents faster. Your work should multiply others, not just your own output.