Mechatronics Scientist

About Maincode

Maincode is an AI research and engineering company and home to Matilda, Australias first and only large language model trained from scratch. We operate our own advanced AI infrastructure in local data-centres; designing, building, and running everything from the racks up through model training and serving systems.

Were building an AI Model Factory, a new kind of compute infrastructure that powers frontier AI research and production systems that work at scale.

About the Role

Despite the title, this role isnt really about traditional mechatronics. Were looking for something that doesnt quite exist in the market: an AI Model Factory Architect.

This role sits at the intersection of physical systems, large-scale compute, and AI research. Youll help design and operate the infrastructure that trains large-scale models: racks of GPUs, high-performance networks, and the orchestration systems that keep them running.

People with strong research instincts and systems thinking thrive here. You dont need prior datacenter experience; curiosity, first-principles reasoning, and the ability to learn fast matter more. Someone from mechatronics, robotics, or aerospace who has strong computational skills could excel in this role. Well give you the time, mentorship, and resources to master how these systems work, and youll develop expertise in one of the most essential technical domains of the coming decade.

You Might Be a Fit If

• You have a background in a technical or scientific field (mechatronics, robotics, aerospace, physics, or similar) and strong computational skills (e.g., modelling and numerical simulation, automation, MPI or HPC environments, GPU and accelerator programming, systems programming, or Linux).

• You enjoy understanding complex systems from first principles.

• You learn new technologies fast and apply them independently.

• You think in terms of system architecture, how components interact and scale.

• Youyoure curious about how AI actually runs, from circuits and cooling to training loops.