Line of Service

Advisory

Industry/Sector

Technology

Specialism

Advisory - Other

Management Level

Senior Associate

Job Description & Summary

The AI team is comprised of highly experienced designers, developers, engineers, data scientists and analysts who are responsible for delivering digital products and AI solutions to PwC’s business across multiple lines of service in Australia.

We partner with subject matter experts across the business to advise on, design and build agentic AI systems and LLM-powered products that transform the way we deliver assurance, advisory, tax and legal services to our clients.

As an AI Engineer on our team, you will design, build and evaluate production agent systems: agents that reason, call tools, retrieve knowledge, manage their own context and complete long-horizon, multi-step tasks on behalf of our practitioners and clients. Our agent harness is built in TypeScript on the Vercel AI SDK, with model and provider routing handled through a LiteLLM-based gateway, and it runs at scale in the Azure cloud. You will own problems end to end: from framing the approach and prototyping, to shipping reliable, observable agents in production.

We are an evaluation-led team, so we care deeply about rigour: measuring agent quality, running evaluations, and using the scientific method to know whether something actually works, not just whether it demos well.

Why PwC?

At PwC Australia, your skills meet purpose. We tackle big challenges across industries like finance, technology, energy, and health—giving you the chance to makeareal impact. Here, your growth is our priority. You’ll work with leading teams, explore new technologies, and unlock your full potential.

Join a global community ofmore than370,000 peoplewhovalue bold ideas, collaboration, and lastingchange. Together, we’re building trust andshapingthe future.

What you’ll do:

- Build and extend our agent harness in TypeScript: the runtime loop, tool dispatch, streaming, retries and tool-call repair, state persistence and recovery, and the guardrails that keep agents reliable and on-budget.

- Design context and memory strategies for long-horizon work: context compaction andsummarisation, working versus long-term memory, prompt and response caching, cache-aware prompt construction, and token budgeting across multi-step runs.

- Architect multi-agent and sub-agent systems: lead and worker patterns, task decomposition, parallel sub-agent execution, hand-off and delegation contracts, shared state, andresultaggregation back to a lead agent.

- Design and expand agent tool ecosystems with well-typed, testable schemas: a sandboxed command tool giving agents a real filesystem and shell, file and document operations, code execution, web search, and connectors that surface enterprise and client data sources (for example MS Graph, SharePoint, Outlook and Teams) to agents as tools.

- Route across multiple models and providers through ourLiteLLM-based gateway, tuning model selection, fallbacks and routing policy for quality,latencyand cost - including the use of small language models for narrow,high-volumeor latency-sensitive steps.

- Engineer retrieval and RAG pipelines that ground agent outputs in trusted sources: chunking, embeddings, vector search on PostgreSQL withpgvector, hybrid and re-ranked retrieval, and retrieval exposed to agents as tools.

- Produce structured, reliable model outputs using typed schemas (for example Zod with theVercelAI SDK) and design robust prompts, systeminstructionsand agent skills.

- Build evaluation harnesses and offline batch evals, including deep-research style runs and trajectory-level assessment of tool use, to measure accuracy, faithfulness,costand latency - and use those signals to iterate.

- Instrument agents with observability and tracing (for exampleLangfuse,OpenTelemetry) and track token usage, cache hit rates and cost in production.

- Produce clean, maintainable,



efficient code deployed at scale in the Azure cloud; scaffold new projects, pair with engineers, and review pull requests.

- Contribute to team stand-ups and the broader software developmentlifecycle, andparticipate in firmwide AI and engineering forums.

What we’re looking for:

Applicants must be able to demonstrate the following key capabilities. We do not expect every box ticked, strong fundamentals and a track record of shipping LLM-powered products matter most.

Core skills:

- Robust TypeScript development experience, ideally in a Node.js backend context, with confident use of the type system to model tool schemas, agentstateand structured outputs.

- Hands-on experience building agents with a modern agent framework or SDK, with theVercelAI SDK strongly preferred (streaming, tool calling, structured generation, provider abstraction).

- Demonstrated experience building oroperatingan agent harness or runtime loop in production: tool dispatch, error handling and tool-call repair, interrupt and resume, and long-running task execution.

- Practical understanding of context engineering: context window management and compaction, memory design, prompt and response caching, and cost/latency trade-offs across a multi-step run.

- Experience designing multi-agent architectures, including sub-agent orchestration, delegation boundaries, sharedstateand failure handling.

- Deep, practical knowledge of prompt engineering, agentic patterns,tooland function calling, and structured outputs.

- Experience routing across multiple models and providers through a gateway such asLiteLLM, and reasoning clearly about modelselectiontrade-offs.

- Experience building andoptimisingRAG and retrieval pipelines, including their evaluation, with industry-standard tooling.

- Strong experience with PostgreSQL, includingpgvectorfor embeddings and hybrid or semantic search, and using Postgres as the durable store for agent state,memoryand traces.

- An evaluation-first mindset: designing evals for agentbehaviour, and using evidence rather than demos to reach defensible conclusions about quality.

- Strong critical thinking, analyticalrigourand outstanding attention to detail.

- Proper source code management and confident use of Git, and comfort working in a large sharedmonorepo(for exampleNxor Yarn workspaces) with project boundaries and shared libraries.

- Good written and verbal communication, and the ability to work effectively with remote teams.

- A proactive, problem-solving approach and the ability to solve complex problems as part of a team.

Desirableskills:

- Experience with small language models (SLMs): selecting, fine-tuning, distilling,quantisingor self-hosting smaller models, and using them for routing, classification,extractionor guardrail steps within a larger agent system.

- Experience with LLM observability and cost/quality tracing (for exampleLangfuse,OpenTelemetry).

- Experience designing sandboxed execution environments for agents:containerisedorvirtualisedsandboxes, shell and filesystem access exposed as a command tool, isolation and resource limits, and safe tool surface design. Familiarity with emerging tool-interoperability standards such as MCP is a plus.

- Experience building connectors and integrations to enterprise systems and data sources, including authentication,permissioning, incremental sync and rate-limit handling.

- Full-stack TypeScript experience (NestJS, Next.js / React) with eithertRPCorGraphQL/ Apollo including codegeneration,



andbuilding interfaces for agent products such as streaming and human-in-the-loop review.

- Experience with computer-use or browser-automation agents (for example Playwright).

- Experience with document generation pipelines (DOCX, PPTX, PDF) or rich-text editor frameworks such as Lexical, ideally over a real-time collaborative layer (Yjs/Hocuspocusor equivalent CRDT tooling).

- Working Python capability, includingbuildingand running production services (for exampleFastAPI) as well as evaluation, dataworkand model experimentation; comfort moving across the TypeScript and Python stack is valued.

- A research background in ML or LLM development, and the ability toidentifyemerging techniques and apply them to practical situations.

- Knowledge of classical ML and deep learning, preferably in NLP.

- Experience with well-structured backend services at scale: microservices andcontainerisation(Docker), and ideally domain-driven and event-driven patterns such as domain / application / infrastructure layering,repositoriesand domain events.

- Experience with durable job queues and background processing for long-running agent workloads (for exampleBullMQon Redis, Azure Service Bus), and with Redis for caching,locksand shared state.

- Experiencedeveloping oncloud environments, particularly Azure (Azure OpenAI, AI Search, Blob, Key Vault, App Insights).

- Knowledge of agile software development lifecycles (SDLC), and exposure to CI/CD pipelines (Azure DevOps, GitHub Actions) andinfrastructure ascode (Terraform).

What you’ll gain:

- The opportunity to work with leading organisations and cutting-edge technology that challenge and expand your expertise.

- Workwithin an inclusive environment committed to Inclusion and Diversityand First Nations Prosperity, ensuring diverse perspectives are at the forefront of our work.

- Flexibility means trusting you to choose when, where, and how you workbalancingwhat’sbest for you, your team, and your clients to grow and thrive together.

- Clear paths for career progression supported by continuous learning and leadership development.

- Generous leave entitlements to support your work-life balance - including floating public holidays, birthday leave and ability to purchase additional leave on top of your four weeks.

- A world-class parental leave policy offering up to 26 weeks of leave for caregivers, supporting your family and personal life.

We know that perks are as important as your financial rewards. Explore all the benefits that PwC has to offerhere.

_We’recommitted to treating all our job applicants fairly and with respect, irrespective of their actual or assumed background, disability, neurodivergence, or any other protected characteristicand tomaintaininga safe, respectful workplace for everyone.We want you to have every prospect to thrive in our selection process. In the application form, you can let us know what adjustments yourequireduring our recruitment process and/or any workplace accommodations youanticipateneeding to help you perform your role.

Education

(if blank, degree and/or field of study not specified)

Certifications

(if blank, certifications not specified)

Required Skills

AI Agents, Data Science

Optional Skills

Accepting Feedback, Accepting Feedback, Active Listening, AI Implementation, Analytical Thinking, C++ Programming Language, Communication, Complex Data Analysis, Creativity, Data Analysis, Data Infrastructure, Data Integration, Data Modeling, Data Pipeline, Data Quality, Deep Learning, Embracing Change, Emotional Regulation, Empathy, GPU Programming, Inclusion, Intellectual Curiosity, Java (Programming Language), Learning Agility, Machine Learning {+ 25 more}

Desired Languages

(If blank, desired languages not specified)

Travel Requirements

Up to 20%

Available for Work Visa Sponsorship?

Yes

Government Clearance Required?

No

Job Posting End Date

September 30, 2026

#J-18808-Ljbffr

📌 AI Engineer | Agent Engineer (Agentic AI) (New South Wales)
🏢 Pwc
📍 New South Wales