About the role

We are a small, specialist analytics business. Statistical and machine learning modelling is not a support function here — it is the primary driver of what the business produces. We need a Chief Modeller who understands that difference.

We are looking for someone to step into a player-manager role: leading and developing a team of three, while remaining a direct and active contributor to the modelling work itself. This is not a role for someone who wants to move fully into management and hand the modelling off to others.

The team works on end-to-end statistical and ML modelling pipelines applied to real-world data: from raw data through cleansing, feature engineering, feature selection, model building, evaluation, and live deployment. The Chief Modeller will be responsible for the quality and integrity of that pipeline, the development of the people working within it, and the long-term technical direction of the modelling practice.

This is a boutique operation. The Chief Modeller will have genuine influence over how we work and what we prioritise — with the accountability that comes with that. There is no large technical team to delegate to and no institutional buffer. The role requires both depth and breadth.

Key responsibilities

- Own the technical quality of the full modelling pipeline — data, features, model structure, evaluation, and live performance
- Establish and maintain standards for model documentation, code review, and modelling practice across the team
- Lead the ongoing development and improvement of the team's codebase, balancing delivery continuity with longer-term code quality
- Provide technical direction on modelling approach, tooling choices,



and methodology
- Review and sign off on model changes before they enter production
- Contribute directly to feature engineering, model development, and evaluation — particularly on higher-complexity problems
- Diagnose and investigate model failures and performance issues in production
- Drive modelling research: identify what is worth investigating, design tests, and evaluate results against known benchmarks
- Lead the development of neural network frameworks as a strategic extension of our existing modelling capability
- Mentor the Senior Modeller and the two Junior Modellers, with structured knowledge transfer across the pipeline

About you

- PhD in a quantitative field is preferred — statistics, mathematics, econometrics. Candidates without a PhD but with a demonstrable track record of applied modelling at equivalent depth will be considered
- Substantial post-PhD experience applying statistical and machine learning methods to real prediction problems — not just academic or research settings.
- Demonstrated depth in the full modelling pipeline: data quality, feature engineering, model building, evaluation, and live deployment
- Strong working knowledge of at least one of: Python, R, Stata, or MATLAB




- Practical experience designing and building neural network models applied to structured, feature-engineered data
- Experience mentoring or leading technical staff — the ability to transfer knowledge, not just hold it
- Intellectual honesty about model performance: comfortable with failure analysis, critical of backtest results, and clear-headed about the gap between a paper result and a live one
- Genuinely interested in prediction problems — not just in deploying models, but in understanding why they work or fail
- Comfortable with the absence of clean answers. Production modelling against real-world data is iterative, partial, and humbling. Work with sports data is highly valued.
- Willing to stay hands-on. The role requires continued direct contribution to modelling work indefinitely

About us We are a boutique specialist analytics firm. The team is small by design — we believe that a small group of people working at a high level produces better modelling outcomes than a larger group working at an average one. The modelling work is the core of what we do, and the quality of that work is what determines our results.

We have a ten-year history of applied modelling in this space. Our IP is valuable, our methods are proprietary, and we take both seriously. We invest in tooling, process, and documentation to make our work reproducible and maintainable.

The team that the Chief Modeller will lead is at an early stage — the two junior modellers are new hires and will require significant investment in the first year. This is a formative period for the team, and the Chief Modeller will have a solid hand in shaping how it develops.

📌 Head of Modelling (Brisbane)
🏢 Cumming Technologies
📍 Brisbane