Scoping and requirements

• Meet with department stakeholders to understand the process, the problem and the outcome they actually need.
• Map current state processes and identify where automation delivers real value rather than automating waste.
• Write clear scoping documents covering requirements, proposed approach, systems and data involved, effort estimate and success measures.
• Recommend the right approach for each project, whether that is a platform workflow, an agentic system, an LLM assisted step or a code solution.
• Push back on requests where automation is not the right answer, and say so with reasoning.

 

Engineering and Delivery

• Build, test and deploy automations end to end, from approved scope through to production and handover.
• Design integrations across our systems using REST APIs, webhooks and file based transfers, including authentication, rate limiting, pagination, retries and data mapping.
• Write production JavaScript and Python for logic that a platform cannot handle properly.
• Build systems that survive contact with reality: idempotent operations, sensible retry and backoff, queueing, dead letter handling, and failure states that alert someone rather than failing silently.
• Model and manage the data behind multi step processes, including state, queues and reconciliation.
• Work with our existing stack, which includes NetSuite, Microsoft Dynamics Navision, Magento 2, Freshsales, Freshdesk, Stripe, Google BigQuery, Twilio and Asana.    

 

Tooling Philosophy

• We use n8n as our primary automation platform and we want to keep it that way. Platform based workflows are visible, supportable and maintainable by more than one person, which matters far more over a five year horizon than elegance does. Your default should be to build on the platform where the platform can genuinely do the job. 
• You drop to code when the platform is the wrong tool, not when code is the more interesting option. When you do write code, it is production code: version controlled, documented, tested and deployable by someone other than you. 

 

AI and Agentic Systems

• Build agentic workflows that combine planning, tool use and decision making using LLMs. 
• Integrate LLMs into business processes for classification, extraction, summarisation, document routing, drafting and decision support.
• Treat prompts as production assets: versioned, tested, and managed with attention to context, structured outputs and tool definitions. 
• Manage agent memory and state across multi step actions, and design guardrails, validation and human review points where the cost of an error is high. 
• Evaluate output quality with a defined method rather than by impression, and monitor it after launch. 
• Control token cost and latency as first class design constraints. 
• Recognise when an LLM is the wrong tool and use ordinary code instead. 

 

Reliability, Documentation and Support

• Own the reliability of what you build, including monitoring, logging, alerting and error handling. 
• Maintain technical documentation and runbooks in Confluence so automations can be supported by others. 
• Track work in Jira and Asana and keep status current without being chased. 
• Apply sound security practice around credentials, API keys, permissions and the handling of customer and commercial data. 
• Support and improve automations already in production, not only new builds. 

 

In addition to the duties listed above, the position holder must carry out tasks assigned by the supervisor that are essentially related to duties.

• Excellent written and spoken English is essential and is weighted equally with technical ability. You will scope requirements directly with English speaking stakeholders in Australia, the United Kingdom and the United States, and you will write the documentation those stakeholders depend on. Strong technical skill will not compensate for weak English in this role.
• Minimum 7 years of professional experience in software development, systems integration or automation engineering. 
• The seven year requirement applies to software engineering, integration and automation work. We do not expect seven years of LLM or agentic experience, because the tooling has not existed that long. AI experience is assessed on depth and recency instead.
• Strong production level JavaScript and Python. You can solve a problem in code rather than being limited by what a platform offers.
• Solid engineering fundamentals: version control and branching in Git, SQL and data modelling, API design and consumption, testing, structured logging, and deployment.
• Working experience with Git hosted in Bitbucket or equivalent, and with Jira, Confluence and Asana or comparable tools. You keep tickets, branches and documentation in order.
• Practical experience with n8n, Make.com, Zapier or an equivalent automation platform, including complex multi branch workflows running in production.
• Deep knowledge of APIs, webhooks, authentication methods and integration patterns, including working against poorly documented or awkward third party APIs.
• Demonstrated experience building LLM and agent based systems that reached production and stayed there. Familiarity with Claude and the Anthropic API is highly regarded.
• Practical prompt engineering experience, including structured outputs, tool use and managing agent state across actions.
• Proven ability to scope work independently with non technical stakeholders and turn a business requirement into a technical solution.

 

Nice To Have

• Experience with ecommerce platforms, particularly Magento 2 and Stripe.
• Experience integrating with an ERP such as NetSuite or Microsoft Dynamics Navision.
• Experience with data warehousing and reporting, including Google BigQuery and GA4.
• Exposure to DevOps practices, CI/CD pipelines or container based deployment.  
• Familiarity with process mapping and opportunity discovery methods such as value stream mapping or SOP analysis.  
• Experience working in a distributed team across multiple time zones.