Lead ML Engineer – Defect Detection & Physical Systems

The Role

We are looking for an autonomous Machine Learning Engineer who is energised by difficult, ambiguous problems.

This is not a clean-data Kaggle competition or an academic research gig. Industrial sensor data is messy, physical processes are non-linear, and edge hardware has real compute limits. As our primary ML lead, you will take end-to-end ownership of our classification and 3D segmentation pipelines. You will investigate anomalous physics, architect resilient deep learning models, write clean production code, and see your work deployed directly onto machines in the field.

If you are someone who gets bored when the answers are in the first page of search results, and instead gets energised when you have to invent or adapt solutions to stubborn, real-world problems, you’ll find your home here.

What You'll Do

• Own the ML Core: Architect, train, and deploy deep learning models for classification, anomaly detection, and segmentation across complex image and 3D volumetric datasets.

• Tackle Unstructured Data: Build robust preprocessing, filtering, and data-augmentation pipelines to extract signal from sensor streams.

• Ship to Production: Collaborate closely with the software team to embed models into our production codebase, profiling and optimising inference for edge and cloud deployment.

• Drive Applied R&D: Continuously scout, evaluate, and benchmark novel architectures that push the limits of real-time defect detection.

• Debug Across the Stack: When a model behaves unexpectedly, you’ll dive deep; inspecting hardware capture logs, raw sensor data, and pipeline code to root-cause the issue.

What You'll Bring

• Proven Problem-Solving Grit: A track record of untangling complex technical or analytical problems; whether in commercial software, research, competitive tech, or deep-dive personal projects.

• Deep Learning Fluency: Deep practical knowledge of modern DL architectures and frameworks (e.g., PyTorch), with hands-on experience in computer vision, image segmentation, or 3D data.

• Solid Software Engineering: Strong proficiency in Python and its scientific stack, paired with solid engineering fundamentals (clean code, Git, Docker, Linux, CI/CD).

• Pragmatic Mindset: The judgment to know when an off-the-shelf architecture solves the problem versus when a custom pipeline is truly warranted.

• Right to Work: You must have valid working rights in Australia.

• Bonus Points: Exposure to edge model optimization (TensorRT, ONNX), spatial computing, AWS infrastructure, or manufacturing/hardware environments.

What We Offer

• Competitive Package: $150,000 p.a. + 12% super.

• Meaningful Equity: A real stake in the company via our employee share scheme.

• Dedicated Growth Time: Paid, protected time for self-directed learning and experimentation to keep you ahead of the ML curve.

• High-Impact Mission: The opportunity to solve novel technical problems with immediate physical consequences across global industries.

• Collaborative Hub: A flexible, pragmatic team environment based in our Oakleigh facility with direct access to physical hardware and 3D printing equipment.

About Additive Assurance

At Additive Assurance, we tackle one of the most demanding problems in advanced manufacturing: ensuring metal 3D-printed components are built right, every single time. Companies in aerospace, defense, and medical devices use 3D printing to create critical parts; from rocket nozzles to orthopedic implants. In these industries a tiny flaw can be catastrophic.

Our leading AMiRIS system inspects the laser powder bed fusion process layer-by-layer, catching anomalies in real time. We operate at the intersection of physics, optical sensing, and computer vision. Your work will directly ensure that advanced manufacturing becomes dependable enough to trust with human lives.

A Note on Our Culture: Will You Thrive Here?

We are a small team of builders and tinkerers who measure success by tangible, deployed impact. We iterate rapidly, test hypotheses against ground truth, and adapt as we learn.

To thrive here:

• You lean into hard technical walls rather than waiting for someone to give you a roadmap.

• You care as much about clean code and edge compute efficiency as you do about model accuracy.

• You value shipping a pragmatic, robust solution into production over perfecting a theoretical model that never runs in the field.

How to Apply: Show Us How You Solve Hard Problems

We don’t just look at standard resumes or keyword lists. We want to see how you think when things get difficult.

Along with your CV, please include a short story (<400 words) about a genuinely tough problem you tackled. It could be anything, for example:

• A stubborn algorithmic bug or training instability you tracked down and fixed.

• A messy personal hardware/software project, open-source contribution, or hackathon challenge.

• An unconventional, creative workaround you built when standard tools failed.

Tell us what broke, why standard approaches didn't work, how you diagnosed the root cause, and how you delivered a solution.
Please upload via the Seek application process.