Senior Machine Learning Engineer

Salary

$150,000 – $180,000 per year

Key Responsibilities

• Set up National Digital's model development practice, including tooling, environments, experiment tracking and review
• Design and fine-tune models for client problems such as document extraction, classification and domain-specific assistants
• Build evaluation suites that run in CI and gate every model and prompt change
• Run models in Australian regions where a client's data must stay onshore
• Work with backend engineers to put models behind APIs with monitoring, guardrails and cost limits
• Follow research and new model releases, and test which of them change what we can offer clients
• Explain model capability and risk to clients in plain language

Requirements

• 5+ years in machine learning engineering, with models you trained or adapted running in production
• Fine-tuning and post-training experience with open-weight language models, including supervised fine-tuning, preference optimisation and parameter-efficient methods such as LoRA
• Evaluation design, including task-specific test sets, graded rubrics, model-graded evaluation and the statistics to separate a real improvement from noise
• Model design judgement about when a prompt is enough, when retrieval is needed, when to fine-tune and when a small classifier will outperform a large model
• Training data work, including collection, labelling, synthetic data generation and quality control within Australian privacy law
• Serving and optimising models in production, including quantisation, batching, latency and cost per request on cloud GPU infrastructure
• Production Python and the PyTorch ecosystem
• A degree in Computer Science, mathematics, statistics or a related field, or equivalent research or industry experience
• Based in Australia with full, unrestricted working rights (Australian citizen or permanent resident)
• Willingness to complete a National Police Check and provide professional references prior to commencement

Bonus Skills / Nice to Have

• Agent systems with tool use, planning and memory, and evaluation of multi-step agent behaviour
• Interpretability or AI safety research experience
• Multimodal models for documents, images or speech
• Published research, or open-source contributions to model training or evaluation tooling
• Experience with AI governance standards such as ISO/IEC 42001 or Australia's Voluntary AI Safety Standard

Contract Details

📝Full-time / Remote (Australia), starting in the first half of 2027⏰38 hrs/week

Benefits

✓Full-time remote work anywhere in Australia, with flexible hours✓$150,000–$180,000 plus superannuation, depending on experience✓A compute budget for training and evaluation work✓The opportunity to set up a new practice inside a small senior team✓A professional development budget that includes conferences