Staff Machine Learning Engineer (Platform)

About the role

Neara builds physics-enabled digital twins of electricity grids using machine learning, helping utilities understand risks and optimise investments. The Staff Machine Learning Platform Engineer owns the infrastructure and systems that let the ML discipline move fast, ship reliably and scale. The role covers the full ML platform strategy, from training pipelines and serving architecture to experiment management and monitoring, and involves deploying frontier spatial models to global utility and new vertical customers.

Key responsibilities

• Own the ML platform strategy end-to-end: define and drive the multi-year technical roadmap for training pipelines, serving architecture, experiment management and monitoring systems.
• Build tooling that accelerates ML delivery, standardising workflows and removing friction between experimentation and deployment.
• Solve distributed systems problems: enable training across distributed data with residency and security requirements, and make models run efficiently across varied GPU hardware, including sparse tensor implementations.
• Design scalable, flexible serving architecture that handles spiky production load while letting the ML team experiment across regions, customers, tasks and verticals.
• Identify what slows the team down, define the contracts and interfaces between training, evaluation and serving, and build the roadmap to turn ambitious research into routine delivery.

Required profile

• Significant hands-on experience building and operating training pipelines, distributed compute, model serving and monitoring systems at scale.
• Proven experience in production model monitoring, data quality frameworks and training data warehouses brought to ML readiness.
• Ability to set ML platform standards, influence engineering roadmaps and drive alignment on complex infrastructure decisions across teams without direct authority.
• R&D foundation that informs good prioritisation decisions.
• Experience leading, mentoring and coaching ML engineers, raising the technical bar through hands-on guidance, documentation and internal standards.

Required skills

• Strong proficiency writing and optimising custom CUDA kernels for deep learning training, ideally on non-text/image data types, and diagnosing performance bottlenecks in sparse architectures.
• Proficiency in Python and PyTorch (or equivalent).
• Experience with AWS, GCP or Azure, container orchestration (Kubernetes, Docker), and on-prem or neocloud environments.

What we offer

• Competitive salary and meaningful ESOP.
• Fully flexible work environment, with a fully stocked office in Redfern and regular office events.