Wellington, New Zealand
Permanent, full-time
As a Software Engineer - AI at Hnry you will help grow our delivery pipeline, which uses agents to take simple, repetitive engineering work from bug report to merged fix. You’ll work within our dedicated AI squad and with the engineering teams across Hnry to build the platform and standards for agents.
We think of our agents, shared skills and the people who use them as one system, and you’ll design for the whole system rather than one agent at a time. The role isn’t a service for building every agent idea the business has. Instead, it makes the agents that Hnry already relies on faster, cheaper, safer and easier to run.
If this sounds like something you’d be keen to be part of, we’d love to hear from you 🫶🏽.
🚀 Responsibilities
Pull common logic out of our most-used agents into a shared skills layer, so large agents become thin orchestrators that call reusable skills.Optimise agent cost and run time, for example by matching the right model to each subagent step and tightening prompts.Track cost and run-time metrics for our key agents, and report the trends so improvements are visible across the team.Run agents on managed hosting, and set up the environments, access, secrets and service accounts they need.Build production readiness into every agent: monitoring, alerting, logging, audit trails and evals (automated checks that catch drops in quality), with security and privacy guardrails for agents that touch customer personal data or support compliance work.Apply our security and privacy guardrails, especially for agents that touch customer personal data or support compliance work.Design feedback loops that make agents better over time: capture what users flag (Notion comments, Slack reactions, overridden outputs), route it to the agent’s owner, and turn it into tested changes to prompts and skills.Respond when agents fail: diagnose the cause, fix it, and share what you learned so the same failure doesn’t recur.Shape how people and agents work together: where a human reviews, redirects or approves, and how skills are built so both people and agents can use them.Build and improve the delivery pipeline for low-complexity, high-frequency work.Hold every agent-assisted change to Hnry’s engineering standards: tested, reviewed, and defensible at a whiteboard. Review code thoroughly and document what you build so others across Hnry can reuse it.🫵 You will:
Own technical work from idea through to production and ongoing support.Build with LLMs or agent tooling such as Claude Code, subagents, skills, MCP or similar.Get an app or agent running on a hosted platform, and learn the rest with support from the team.Keep cost and observability front of mind, using tools like Datadog, Rollbar and BigQuery.Weigh model trade-offs (cost, speed, quality) and test systems whose output isn't always the same.Handle secrets, access and personal data with care.Think in systems: see how agents, shared skills, feedback and the people using them fit together, and spot where one change improves the rest.Ship small, measure, and iterate.Explain technical trade-offs clearly to technical and non-technical people.Bonus points for the following:
Experience building large scale systemsExperience being self-employed (freelancing, contracting etc) in any industry/professionExperience using Ruby on Rails💪 How we work
We follow agile practices, delivering improvements iteratively in small chunks. We track the impact of our work and measure ourselves on results, not the amount of code written.AI is part of how we work at Hnry. It’s been adopted quickly across the business to cut down on admin and efficiency work, and we’re building our own AI agents to make internal processes and information easier to access. Judgement-based work stays with people.We’re growing fast, and are focused on building a well designed, scalable, and robust platform. In some areas, such as customer data and compliance, there is very little margin for error.Our technology platform is built in Ruby on Rails. We also use a range of AWS services (managed via Terraform), Postgres, Docker, and a little React.We have automated tests, continuous deployment, and release to production multiple times per day.We review each other’s code and set aside time for experimenting with new features, refactoring, and learning new things.We work hard and finish on time, no crazy hours.unnamed (10).png
Salary banding: $90,000 - $110,000 per annum
✨ At Hnry, we operate a policy of pay parity across our organisation to ensure that all Hnry staff are fairly and equally compensated based on their contributions and experience. The salary for this role will be determined based on the applicant’s experience✨
💜 How to apply
Complete the Hnry Job Application Form and we will be in touch!