SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Lumi is a leading Australian fintech small business lender that has funded over $1.5 billion since 2018, serving 15,000+ small businesses. They are seeking an AI Lead to architect and scale agentic AI workflows across the organization.
In this role, you will partner directly with business and product teams to identify, scope, and build agentic workflows, AI agents, and reusable skills that solve real problems. You'll advise on platform decisions for where automations and agents are hosted (e.g., n8n, Databricks Apps, Claude-based tooling), balancing speed of delivery against maintainability, cost, and governance requirements.
Key responsibilities include:
- Designing how agents and automations access data, determining the right query layer (governed semantic layer vs. raw source), permissions, guardrails, and avoiding shadow pipelines
- Shaping data organization to support automations by partnering with Data Engineering on cataloguing, metadata, and lineage (e.g., via OpenMetadata)
- Establishing ownership and stewardship models with clear domain stewards and automation owners to maintain accountability post-launch
- Running governance scaffolding for AI/automation adoption—tiered autonomy classification, risk assessment, and lifecycle management (build → test → deploy → monitor → retire)
- Working closely with the Data Architecture & Strategy Lead to prioritize which processes get automated against strategic initiatives
This is a new kind of role at Lumi, and you will help define its shape as much as fill it.
REQUIREMENTS:
- Strong prior experience shipping production software
- Hands-on experience building agentic workflows and automations in tools like n8n, and working with governed data platforms such as Databricks
- Practical experience with agent patterns: tool use, function calling, retrieval-augmented generation, multi-step orchestration, structured outputs
- Hands-on experience with LLM evaluation and observability/monitoring tooling (e.g., LangSmith, Langfuse, Braintrust)
- Systems mindset—able to think simultaneously about the automation itself, the data it depends on, the platform it runs on, and accountability for each piece
- Experience or strong interest in data governance and metadata management (e.g., OpenMetadata, data cataloguing, lineage) and how it underpins reliable automation
- A builder's instinct paired with a governance instinct—comfortable shipping quickly while also stepping back to ask about ownership, accountability, and guardrails
- Strong stakeholder skills—able to translate manual or ad hoc processes into well-scoped automations and bring the right people into ownership conversations
- Comfort operating with ambiguity in a newly defined role
- Experience in fintech, lending, or financial services highly regarded
