AI/ML Career
9,582 roles at 2,821 companies

Senior Software Engineer - Internal AI Accelerator Squad

The role / impact Every team at Xero has work that is slow, manual and error prone and the people doing it are often the ones best placed to make better decisions if they had time. We're building agentic AI workflows/agents that take that toil away, so our people can focus on the work that matters. The internal AI accelerator squad partners directly with Xero's enterprise and G&A functions, including HR, Legal, Finance, Sales teams. As a Senior Engineer, you will work in a forward deployed way, sitting down with senior business leaders to understand their challenges, then designing, building and delivering agents and workflows that the business can run itself. You will own that work end to end, from the first conversation to a live system with real users. The technical challenges: The Squad is early in it's life and the problems are as much about judgement as engineering. Five problems define the work: - Agents the business trusts: Build agentic workflows that handle real work, with human review, guardrails and clear escalation, so business teams are comfortable relying on them. - Discovery to delivery: Take an ambiguous business pain point, shape it into a problem worth solving, and deliver a working solution quickly, then iterate with the people using it. - Production, not prototypes: Design agents that are observable, evaluated and maintainable, with evals, monitoring and cost controls built in from the start. - Handover that sticks: Build so the business team can own and run the outcome without you in the loop, and so the next agent is faster to build than the last. - Platforms that scale us: Build the golden paths, MCP integrations and reusable skills that let hundreds of Xeros build and ship AI safely, without needing an engineer for every idea. What we build: - Agents in production: triage agents for IT and People support, an onboarding orchestrator that guides every new starter, tax classification, and spend limit approval automation that cut approvals from hours to seconds. - Platforms that scale AI across Xero: the Dashboard Accelerator, which takes apps people build in Claude to secure internal hosting in minutes instead of weeks; a governed marketplace for Claude plugins and skills; and support for a growing community of citizen builders. - MCP and enterprise integrations: production grade MCP servers that safely connect agents to systems like Workday and Salesforce, plus the patterns and guidance the rest of Xero builds on. - Bonus: experience building MCP servers, durable workflows (e.g. Temporal), or agent frameworks such as Pydantic AI or LangGraph. Tools and Tech: You'll get early access to the latest AI tooling (Claude Code, Glean, Gemini, GitHub Copilot and many models through our AI gateway) and the freedom to choose the right tool for the job. Today we build Python, Typescript and .NET; MCP;Temporal; AWS, Kubernetes and Terraform; Snowflake and ddbt; and GitHub Actions. Your job is to keep adding to that list, with each workflow sharper and more reusable than the one before. Impact: - First 6 months: You'll understand the squad's platform, patterns and stakeholders. You'll be building alongside the team and have one agentic workflow in discovery or build with a senior business partner, with a clear scope and success measures agreed up front. - First 12 months: You'll have delivered at least one agentic workflow into production that a business team uses day to day, with evals and monitoring in place. You'll have handed it over so the business runs it, and your learnings will be feeding back into the squad's shared patterns and tooling. You will have also built a strong rapport with the business function you are working with, being the trusted technical expert for all things AI and contributed to the squads shared platforms, integrations or tooling. What You Bring - Software engineering fundamentals, with experience designing and building reliable services and APIs - Experience delivering agentic AI or LLM workflows into real production use, not just prototypes or personal projects - Experience owning a relationship with senior non-technical stakeholders: understanding their pain point, building a solution, and handing it over so they run it without you - Experience designing agentic systems with tool use, retrieval, human-in-the-loop controls and guardrails, and evaluating them with measurable quality criteria - The ability to explain technical trade-offs clearly to non-technical leaders and build trust through delivery - Comfort using AI tools to improve how you build, with judgement about when and how to validate their output How We Work We're bold in our vision and pragmatic in execution. We deliver useful improvements early and often, and we invest in AI to learn faster and focus effort on higher value problems. Engineers stay accountable for what they ship, including quality and reliability. We use shared frameworks to avoid reinvention, leaving teams room to innovate. Leaders set context and priorities, and engineers help shape the solutions. Where we work: We offer a hybrid model that blends the flexibility of working from home with the collaboration of the office including designated boost days in our office spaces. Our hiring process: 1. TX phone screen (45 minutes): a conversation with your Talent Specialist about your background, motivations and the role. 2. Round 1: (60 minutes) A technical conversation with Senior Engineers on this team, where you'll walk through an agentic workflow you've delivered, from the business problem to production and core engineering technical questions 3. Round 2: (60 minutes) A conversation focused on how you work with stakeholders and show up against Xero's values, including ownership, customer thinking and AI fluency. If this sounds like the role you are looking for, hit the apply button!

Summary

Senior engineer on an internal AI accelerator squad, building agentic AI workflows for Xero's enterprise and G&A functions end to end, from discovery with senior business leaders to production handover. Requires experience delivering agentic AI or LLM workflows into production and designing systems with guardrails, human-in-the-loop controls and evaluation. Tech includes Python, Typescript, .NET, MCP, AWS and Temporal.

Responsibilities

Partner with enterprise and G&A functions to design, build and deliver agentic AI workflows end to end; build trusted agents with review, guardrails and escalation; ensure handover so business teams run outcomes themselves; build shared platforms, MCP integrations and reusable skills

Qualifications

Experience delivering agentic AI or LLM workflows into production; experience owning relationships with senior non-technical stakeholders; ability to explain technical trade-offs to non-technical leaders; comfort using AI tools with judgement about validation

Experience requirements

Experience designing and building reliable services and APIs; experience designing agentic systems with tool use, retrieval, human-in-the-loop controls and guardrails

Interview process

TX phone screen (45 minutes); Round 1 (60 minutes) technical conversation with Senior Engineers; Round 2 (60 minutes) stakeholder and values conversation