Agentic Software Engineer II

Agentic Software Engineer II

Set up, direct, and ship with AI coding agents as core tools, own every part of what you build, and reach millions of travelers.

As an Agentic Software Engineer II, you are a product-minded, full-stack engineer who owns complex features and small AI workflows from idea to production. You treat AI coding agents as systems to set up and direct, not just autocomplete, and you bring the judgment to review, correct, and stand behind what they produce.

This role supports the technology foundation behind arrivia, a leader in travel loyalty, membership, and cruise solutions. You will decompose the work, anticipate dependencies, negotiate realistic scope, and ship on time with minimal supervision.

The features you ship reach millions of members through some of the world's most recognized brands. Their trust is the product, so your decisions on approach, quality, and safety shape delivery timelines, code maintainability, and the experiences members remember.

This is a builder's seat with real room to grow. You will set up context, plans, and checkpoints for agentic tools and guide them through real work, choose the right models and architectures for each problem, design mitigations for common AI failure modes, and upskill the engineers around you. It is a fast-paced, collaborative environment that spans multiple technology domains and teams.

• Ship end to end: Own delivery of complex features and small AI workflows for our travel platform, from decomposition through production, keeping launch checklists and SLOs healthy.
• Direct AI agents: Set up context, plans, and checkpoints, then guide agentic tools like Claude Code, Cursor, and Git Hub Copilot through real work, verifying each step rather than accepting whatever appears.
• Own what you ship: Review AI-generated and hand-written code for correctness, security, and maintainability, and be able to explain every part of what you deliver, whether you or an agent wrote it.
• Choose the right approach: Select appropriate models, tools, and architectures for each scenario, anticipate common AI failure modes, design sensible mitigations, and know what to keep by hand.
• Practice responsible AI: Apply guardrails, content filters, and grounding patterns, and raise concerns when outputs could harm members, violate policy, or mislead.
• Extend agentic workflows: Use and improve Model Context Protocol (MCP) integrations and internal agentic workflows that connect AI agents to our tools and services.
• Measure what you build: Define success metrics and instrumentation for your features, and design simple experiments with clear hypotheses using feature flags and prompt variants.
• Protect member trust: Safeguard the privacy and security of members' sensitive data, in partnership with our Risk and Security teams.
• Stay close to members: Engage with customer feedback, support tickets, and user research to refine your solutions, and push back when a request does not align with real member value.
• Lift the team: Onboard and upskill newer engineers on core tools, review their work helpfully, and model good craft.

• 2 to 5 years of full-stack development experience building web applications and services.
• A recent, hands-on track record. You have personally built and shipped working software in the last year, and you can explain how you built it.
• 1+ year of hands-on experience using AI coding agents such as Claude Code, Cursor, or Git Hub Copilot in professional workflows, and a way of working that goes beyond back-and-forth chat.
• Proficiency in one or more of Type Script/JavaScript, Python, or C#/.NET.
• Experience with modern frontend frameworks such as React, Next.js, Vue, or Angular.
• A solid grounding in RESTful API design and microservices patterns.
• Experience building and deploying applications on a cloud platform (AWS, Azure, or GCP).
• Working knowledge of prompt engineering for code generation, and familiarity with the Model Context Protocol (MCP).
• Bonus points for strength in any of these:
  Retrieval-Augmented Generation (RAG), Lang Chain, Azure AI Foundry, graph-based data (GraphRAG, GraphDB), streaming or messaging platforms…