Senior Software Engineer, AI Development
Location: Australia (remote-first, AU-based)
Employment type: Full-time
Salary: Competitive senior market rate
The problem space
Refactor are partnering with an Australian AI startup with a production AI platform, enterprise customers and a growing SaaS offering.
We’re now extending the platform to support an LLM-agnostic, agentic AI assistant that integrates across enterprise systems and helps everyday business users plan work, prioritise tasks, and structure their work experience.
This role exists because the system is no longer experimental. Retrieval quality, orchestration, latency, security and UX are all priorities in our ongoing product development.
What we’re building
The platform is built around production-grade RAG and agentic workflows.
Backend:
A Java-based API platform that:
- Orchestrates LLM calls
- Executes multi-step agent workflows
- Integrates with enterprise systems
- Enforces tenancy, security and access controls
Retrieval layer
- Weaviate as the primary vector database
- Hybrid search (vector + keyword)
- Tuned chunking, embeddings and reranking
Frontend:
- React application used daily by non-technical business users
- Surfaces agent outputs, plans, tasks and recommendations
AI layer
- LLM-agnostic (supporting models from Anthropic, Mistral & Open AI)
- Support for cloud, private and sovereign deployments
- Explicit cost, latency and reliability controls
The assistant is agentic by design. The planning, sequencing actions, querying multiple systems, and producing structured outputs users can act on.
What you’ll be responsible for:
You’ll be a hands-on senior engineer contributing across backend and frontend, with a primary focus on core product development.
Backend & product engineering:
- Strong T-shaped engineering skills with deep expertise in backend Java development and practical knowledge across related technologies required to deliver production-ready systems
- Experience with AWS and cloud technologies, CI/CD pipelines, system configuration, unit and manual testing, and integrations with external systems such as web crawlers and enterprise API services
- Developing LLM orchestration and agent execution layers
- Implementing and maintaining enterprise integrations (documents, tasks, calendars, knowledge systems)
- Ensuring APIs are versioned, stable and suitable for enterprise time horizons
RAG & retrieval systems:
- Designing and evolving production RAG pipelines
Owning:
- Document ingestion and preprocessing
- Chunking strategies and embedding lifecycle management
- Retrieval tuning, hybrid search and reranking
- Experience in database-level optimisations, query and retrieval performance tuning, and embedding lifecycle management to improve accuracy, consistency, and efficiency in production datasets
- Improving grounding, relevance and consistency across real customer datasets
- Handling retrieval failure modes and partial or stale data scenarios
System quality, scale & reliability
- Optimising latency, throughput and cost across retrieval and model calls
- Implementing observability (structured logging, metrics, tracing)
- Building guardrails, timeouts and fallback behaviour for agent workflows
- Expertise in observability, structured logging, metrics, tracing, and automated testing strategies to ensure robust, scalable, and reliable production systems
- Contributing to tenant isolation, data boundaries and security controls
Frontend contribution (important, but secondary):
- Practical understanding of frontend requirements, enabling close collaboration with UI teams to align API design and deliver fully integrated, production-ready features
- Translating backend capabilities into clear, usable interfaces for non-technical users
- Contributing directly to the React codebase where appropriate
- Ensuring tight alignment between API design and frontend needs
React experience is not required, but practical exposure is a strong advantage.
Product-driven engineering
- Translating real user workflows into concrete product features
- Making pragmatic trade-offs between accuracy, explainability, performance and UX
- Iterating based on production feedback, not assumptions
What we’re looking for
We’re looking for engineers who have already built and managed sovereign AI deployments in production environments.
Required:
- 5+ years professional software engineering experience
- 2+ years building RAG-based systems used in production
- Strong backend experience
- Hands-on experience with:
- Vector databases
- Embeddings, chunking and retrieval strategies
- LLM orchestration (framework-based or custom)
- Experience supporting production, customer-facing systems
Valued:
- Experience with retrieval quality and failure modes
- Experience with agentic workflows beyond simple chains
- SaaS or multi-tenant architectures
- React experience or strong API-first frontend collaboration
- Enterprise or regulated environments
- AWS, private cloud or hybrid deployments
Why this role might be interesting
- You’ll work on real RAG and agentic systems, designed for scale. You’ll be working at the cutting edge as open source models mature, helping evaluate and deploy the best-fit models responsibly in private enterprise environments
- You’ll be involved with evolving the backend architecture and user-facing product
- Customers are already live, so your work has immediate impact
- Small, senior geo-distributed team with high autonomy and professionalism
- Australian company building globally relevant enterprise AI
When you apply, we’re interested in:
- What RAG systems you’ve built or operated?
- Where they struggled or failed?
- How you approached trade-offs in retrieval, latency and UX?
- What kind of AI product you want to help build next?
This role is a fantastic opportunity to join a growing start up, who have already landed some major accounts and now ready to expand. You'll join an exisiting team, who is increibly talented.
What are you waiting for - get applying now!
James Farrey
Founder & Director
james@re-factor.com.au
#SCR-james-farrey