Lead AI Engineer, Agentic Systems

Minter Ellison; Sydney; On-site; AUD 140,000 - 210,000; Full time

Job summary

MinterEllison is seeking an experienced AI engineer to build production AI systems for a complex, regulated legal environment. You will own multi-agent orchestration, evaluation, deployment and governance interfaces used by legal experts, translating complex judgments into robust technical workflows.

You will work with partners and subject-matter experts on Bayesian reasoning, probabilistic modelling, and OpenTelemetry tracing, while ensuring security, observability and regulatory compliance

Qualifications

• production experience with LLM agent frameworks and tool-calling architectures.
• rigorous evaluation discipline and ability to measure model/system changes.
• experience applying Bayesian inference to real-world problems.
• strong Python software-engineering capability.
• React and TypeScript experience for expert-facing interfaces.
• production AWS experience including security and identity considerations.
• sound judgement when working with non-technical experts.
• understanding governance requirements for AI in regulated environments.

Responsibilities

• design and build multi-agent orchestration and Model Context Protocol servers for production legal workflows.
• create deterministic evaluation harnesses and regression suites to measure agent quality.
• implement OpenTelemetry tracing to improve observability of agent runs.
• model regulatory reasoning through Bayesian networks and probabilistic techniques.
• develop expert-facing interfaces for legal specialists to inspect and edit regulatory rules.
• deploy and operate secure, observable platforms on AWS.
• work with partners and subject-matter experts to translate legal judgement into robust technical systems.
• contribute to governance and assurance for AI used in regulated work.

Skills

LLM agent frameworks

Tool-calling architectures

Evaluation discipline

Bayesian inference

Python programming

React

TypeScript

AWS security & IAM concepts

Governance knowledge

Tools

AWS

OpenTelemetry

Job description

Select how often (in days) to receive an alert:

MinterEllison is one of Australia's largest independent law and consulting firms. With a heritage of almost 200 years, we are known for our excellence, and our authentic and enduring relationships with our clients, our people and our communities. Clients seek out MinterEllison to help them solve complex problems every day. We do this by our full-service legal offering and complementary consulting offering.

We offer opportunities to work on industry-leading mandates for top-tier clients, while being part of a high-performing and collaborative team that values excellence, diversity of thought, curiosity, and inclusion. We value our people and empower them to achieve their ambitions - with the support, trust and autonomy to grow their careers in meaningful ways.

At MinterEllison, we're leading the way with AI and other emerging technologies, powering innovation right across our firm. You will have the opportunity to engage with AI tools we've built, and other external AI tools, to enhance efficiencies and excellence internally and with our clients.

Our AI Enablement team builds agentic AI systems for use in legal work. Its platforms support areas including large-scale due diligence, regulatory reasoning across FIRB, continuous disclosure, competition and environmental law, and visa eligibility assessment. Several systems are already operating in production. You will join a small, senior team working directly with partners and subject-matter experts. The team brings together AI engineering, probabilistic modelling, production platform delivery and practical legal expertise to create systems that experts can inspect and govern.

Your Role

This is an end-to-end engineering role for someone who wants to build production AI systems in a complex, regulated environment. You will take ownership across agent orchestration, evaluation, deployment and the interfaces through which legal experts govern the systems they use. The breadth of this role extends from research-grade modelling to production platform engineering. It offers the opportunity to work with live systems, real users and the experts whose knowledge those systems are designed to encode.

In this role, you will:

• design and build multi-agent orchestration, Model Context Protocol servers and supporting tooling for production legal workflows


• create deterministic evaluation harnesses and regression suites that make agent quality measurable


• implement OpenTelemetry tracing to improve the observability of agent runs


• model regulatory reasoning through Bayesian networks and related probabilistic techniques


• develop expert-facing interfaces that allow legal specialists to inspect and edit regulatory rules


• deploy and operate secure, observable and supportable platforms on AWS


• work directly with partners and subject-matter experts to translate legal judgement into robust technical systems


• contribute to appropriate governance and assurance for AI used in regulated work.

About You

We welcome applications from experienced AI engineers who can demonstrate:

• production experience with LLM agent frameworks and tool-calling architectures, including systems deployed beyond the prototype stage


• a rigorous evaluation discipline and the ability to demonstrate whether a model or system change has improved performance


• experience applying Bayesian inference, Bayesian networks or comparable statistical reasoning to a real-world domain


• strong Python software-engineering capability


• sufficient React and TypeScript experience to build usable, expert-facing interfaces


• production AWS experience, including security and identity considerations relevant to regulated data


• sound judgement when working with non-technical experts


• a strong understanding of governance requirements associated with AI in regulated environments.

Exposure to AI assurance standards such as ISO 42001 will be viewed favourably