Software Engineer (Model Training)

Software Engineer (Model Training)

Melbourne | Full-time | 3–5 Years Experience

The Opportunity:

UCentric is exclusively partnering with a high-growth, Melbourne-based technology venture as they enter a pivotal expansion phase following a successful platform rollout. This organization is at the forefront of Australian-led innovation, now focusing on scaling its engineering capability to support complex product delivery and production readiness.

We are seeking a Software Engineer (Model Training) who is eager to contribute to a live, high-performance production environment. This is an ideal role for an engineer with 3–5 years of experience who possesses strong Computer Science foundations and a genuine passion for engineering excellence.

Key Responsibilities:

• Design and maintain pipelines for training large-scale models across massive clusters.
• Build high-performance ingestion and preprocessing systems for massive datasets.
• Develop tooling for checkpointing, reproducibility, and experiment tracking.
• Debug and monitor long-running training jobs, diagnosing failures across infrastructure and training loops.
• Improve training throughput across compute, memory, and data pipelines.
• Work closely with researchers to translate theoretical ideas into stable production training runs.

About You:

• Experience: 3–5 years of experience building production-grade software.
• Tech Stack: Strong proficiency in Python and exposure to ML frameworks like PyTorch or JAX.
• Systems Mindset: You enjoy debugging complex distributed systems and prefer understanding "how the engine works" over relying on abstractions.
• Precision: You have the patience to diagnose failures that may only appear hours or days into a run.
• Curiosity: A genuine desire to learn how large-scale AI training actually happens at the infrastructure level.

Behaviours/Qualities Who Will Thrive:

• High-stakes training runs require a methodical approach.
• You want to understand the "why" behind system behavior.
• You take pride in the stability and reliability of the code you ship.
• You are calm and methodical when production systems behave unexpectedly.

How to Apply: If you are ready to work on the systems that train the next generation of AI, we want to hear from you.

Please include a link to your GitHub or a portfolio of your technical projects within your CV.

Note: This is a full-time role based in Melbourne. At this time, we cannot offer visa sponsorship; applicants must have existing and unrestricted work rights in Australia.