Senior Machine Learning Engineer (Research Optimisation)

TL;DR

Senior Machine Learning Engineer (Research Optimisation): Bridging the gap between research and production by hardening experimental models and optimizing training performance with an accent on PyTorch workloads and GPU efficiency. Focus on building scalable infrastructure, shared libraries, and robust CI/CD pipelines to accelerate the research-to-user loop.

Location: Must be based in Sydney, Australia. The role follows a hybrid work model requiring in-person collaboration at the Sydney campus.

Company

Canva is a global design platform empowering the world to create, with a mission to make design accessible to everyone.

What you will do

• Productionise research models by refactoring, testing, and containerising code for scalable deployment.
• Profile and optimise PyTorch training jobs to improve GPU utilisation and reduce compute costs.
• Develop shared libraries, SDKs, and inference services to standardise model integration across teams.
• Build and maintain CI/CD workflows, artifact management, and multi-variant rollout frameworks.
• Implement observability, monitoring, and reliability practices across training and inference workloads.
• Collaborate with researchers and platform engineers to remove bottlenecks in the training stack.

Requirements

• Must be based in Sydney, Australia.
• Strong software engineering fundamentals with excellent Python skills.
• Proven experience shipping ML systems in production environments.
• Hands-on experience optimising PyTorch training or inference and profiling workloads.
• Proficiency in containerised environments and Kubernetes concepts.
• Ability to refactor research code into clean, production-grade abstractions.

Nice to have

• Experience with distributed training frameworks like FSDP, DDP, or DeepSpeed.
• Familiarity with model-serving tools such as ONNX, Triton, or TensorRT.
• Knowledge of high-performance storage systems like Weka, Vast, or Lustre.
• Experience with MLOps practices and experimentation platforms.

Culture & Benefits

• Equity packages to share in company success.
• Inclusive parental leave policy for all parents and carers.
• Annual Vibe & Thrive allowance for wellbeing and office setup.
• Flexible leave options to support personal recharge and work-life balance.
• Collaborative, thoughtfully designed campus environment in Surry Hills.

Hiring process

• Interviews are conducted virtually.
• Includes interactive, real-time technical challenges reflecting actual work.
• May involve solving problems using AI tools to demonstrate technical approach.