Seeing What’s Possible
Jobs with Arena
The Arena family of companies is growing and offers exceptional career opportunities within Arena firms and with our portfolio companies. See what's possible for you in the postings below.
Senior Computer Vision - Teach Machines to Spot Danger Before It Happens
Optic Security Group
Software Engineering, IT
Posted on Jun 5, 2026
Senior Computer Vision / ML Engineer, AI that saves lives
inviol. Grafton, Auckland. Full time. On-site / hybrid.
🔥 The opportunity
Drive the machine learning behind AI that saves lives. inviol is one of New Zealand's fastest growing AI-native startups. Our computer vision runs 24/7 on real industrial sites: edge devices processing video on the wall, and bigger models in the cloud for the harder problems. It spots safe and unsafe behaviour in real time, so our customers can coach their teams and build a better safety culture. Fewer injuries, real lives saved.
We're after a senior engineer to take on the computer vision side of our stack, end to end: data, training, models, evaluation, and production. Working with our engineering lead and the wider team, you'll help shape the technical direction for vision, with the autonomy and support to build it. You'll work across the whole toolbox, from lightweight detectors on edge hardware to fine-tuned LLMs and bespoke models for the tough stuff. You don't need to have done every piece before, you just need the judgment to work it out.
We're a tight-knit team, and we run on six principles:
Authenticity — say what's real, even when it's uncomfortable.
Iterate fast — test, learn, and don't be afraid to fail.
Delight customers — take their side and follow through.
Keep it simple — in the product, the comms, and the process.
We're Lego — solve problems together, egos at the door.
Robots are our friends — automate it before you do it by hand.
📌 What you'll do
Own and evolve our data pipelines: ingestion, labelling workflows, curation, versioning, augmentation.
Design and run training pipelines, including experiment tracking, reproducibility, and hyperparameter management.
Develop and fine-tune models across the size spectrum: efficient architectures for edge devices, and larger bespoke or fine-tuned models for harder problems.
Build rigorous evaluation: dataset slicing, regression suites, drift detection, and the metrics that actually predict field performance.
Operate the ML in production: monitoring, retraining cadence, rollout, rollback. AIOps for vision.
Collaborate with the wider engineering team on the Python and C++ inference runtime and edge deployment.
Stay close to customers and field data. The gap between benchmark and reality is where we earn our keep.
🚀 What we're looking for
You're a senior engineer who has shipped real computer vision, and you care how it holds up in the field, not just on a benchmark. You won't tick every box below, and that's fine. If you've built and shipped vision systems and you love this kind of problem, we'd love to hear from you.
A strong track record shipping computer vision or ML systems to production, not just notebooks.
Deep experience with modern CV: detection, segmentation, tracking, pose, classification, plus the training frameworks (PyTorch and friends) and the tooling around them.
Strong, production-quality Python. C++ is a bonus, not a must — we have a C++ stack you can grow into.
Comfortable across the model size spectrum: quantisation, pruning, distillation, ONNX for edge, and fine-tuning larger models for the heavier end.
MLOps and AIOps experience: training infrastructure, experiment tracking, model registries, deployment, monitoring.
Comfortable in Linux, bash, Docker, and Azure, or willing to get there fast.
Pragmatic. You can tell the difference between a model that benchmarks well and one that actually works in the field.
Good judgment about when to reach for an off-the-shelf model, a fine-tune, or something bespoke.
✨ Nice to have
Experience deploying to embedded or edge accelerators.
Strong C++ for the inference runtime and edge deployment.
Background with synthetic data, active learning, or semi-supervised approaches.
Experience with multimodal models or VLMs for vision tasks.
Exposure to safety-critical or regulated environments.
🎉 Why join inviol
Real influence over a core part of the product. You'll help guide the direction and propose the solutions, with autonomy inside a small team that ships.
Within a couple of years we're already working with household names like Coca-Cola, NZ Post, PlaceMakers, Fletchers, Vulcan Steel, Woolworths and Foodstuffs.
An AI-native company: the best tools, and the freedom to automate the boring stuff.
A direct line to customers and to the real-world impact of your work.
Flexible working, modern hardware, and a high-energy office in Grafton, Auckland.
A team that takes the work seriously without taking itself too seriously.
✉️ How to apply
Send us your CV and a short note on a CV or ML system you've taken from data through to production. What worked, what didn't, and what you'd do differently.
Full time, based in Grafton, Auckland.