AI Staff Engineer
- Lead the architecture and end-to-end delivery of complex AI systems, balancing resilience, security, scalability and cost.
- Build agentic AI capabilities, including orchestration, state and context management, evaluation, observability and safety controls.
- Provide technical leadership and mentorship while collaborating with product, design and architecture partners.
We're currently looking for an AI Staff Engineer to join our financial services consulting client in Sydney on a permanent basis to lead the design & delivery of complex, AI initiatives for their clients.
The role will encompass both hands-on engineering & leadership; you will lead technical delivery from initial system design through to development and production, while mentoring more junior devs and enhancing the team’s overall capability and productivity.
Key Responsibilities
- Architectural Strategy & Vision: Lead the design of complex software systems, advocating for a "target state" architecture that prioritizes resilience, security, scalability, and cost-effectiveness across the domain.
- Technical Leadership & Execution: Drive the end-to-end development of innovative solutions to address complex technical challenges, ensuring all implementations align with business needs and engineering best practices.
- Mentorship & Collaborative Culture: Provide high-level technical guidance and mentorship to engineers, fostering a culture of knowledge sharing and collaboration within your squad and across upstream/downstream partners.
- Cross-Functional Partnership: Collaborate with product managers, designers, and architects to translate business requirements into exceptional digital experiences, bridging the gap between technical constraints and product goals.
- Engineering Excellence & Innovation: Champion the evolution of engineering practices and the adoption of emerging technologies to continuously improve team efficiency, system quality, and operational scalability.
Required Skill Set
- Agentic Architecture & Control Patterns: Proven experience building agent loops (ReAct, Planning/Decomposition, Reflection, Multi-Agent Orchestration) and handling non-deterministic system design.
- Tooling & API Design: Experience defining strict JSON/OpenAPI schema interfaces optimized for LLM function calling and self-correcting error handling.
- State & Context Engineering: Expertise in state machine design, long-term memory (Vector/Graph RAG), and dynamic context window optimization (summarization, token budget management).
- Eval-Driven Development (EDD): Hands-on experience building automated evaluation pipelines (combining unit tests, schema checks, and LLM-as-a-Judge) to benchmark agent accuracy and prevent regressions.
- Agent Observability & Tracing: Proficiency with deep trajectory tracing tools (e.g., LangSmith, Phoenix, OpenTelemetry) to monitor execution graphs, latency, and token costs.
- Human-in-the-Loop (HITL) & Safety: Track record of designing HITL approval controls, sandboxed code execution (Docker/E2B), and guardrails against prompt injections or unintended tool actions.
- Technical Leadership & Pragmatism: Ability to define SLAs for non-deterministic products and know when to use deterministic code over fragile open-ended agent loops.
...
We vet all employer accounts and do our best to keep job ads safe, but scams can still occur. Be cautious when sharing personal information — never provide financial details or make payments during the application process. For extra security, use the Apply button on our site when proceeding.