About Mutinex

Mutinex is an AI-powered growth platform used by 100+ global brands. Our measurement technology has processed over $9B in media spend and consistently outperforms legacy approaches.

We’ve spent years solving one of marketing’s hardest problems: knowing what actually works. Our product, GrowthOS, combines advanced data science with intuitive software to help marketers understand the impact of past decisions and make better choices about future spend.

The role

You’ll sit within Model Scale and help ensure our models produce accurate, reliable and decision-useful outputs for every customer. You’ll apply our modelling technology to new and refreshed datasets, monitor models in production, investigate unexpected behaviour, and improve the automated tests and diagnostics behind model delivery.

This is not a role where analysis ends with a notebook. You’ll work across the full path from messy customer data to production model outputs—partnering with Data Science on methodology and feature engineering, Engineering on platform reliability and automation, and customer-facing teams on the context behind the data.

The role suits someone who enjoys combining analytical depth with practical problem-solving. You’ll need the statistical judgement to evaluate model behaviour, the technical skills to investigate data and automate repeatable work, and the operational discipline to deliver consistently in a probabilistic environment.

What you’ll do

- Deliver reliable customer models. Apply Mutinex’s modelling technology to new and refreshed customer data, maintaining a high bar for accuracy, consistency and timeliness.





- Monitor model quality. Evaluate model inputs, outputs and performance in production, identifying anomalies, drift and emerging quality risks before they affect customers.

- Diagnose issues systematically. Connect unexpected model behaviour to changes in data, configuration, implementation or methodology—and use sound judgement about when to fix, collaborate or escalate.

- Automate repeatable work. Build reusable data checks, model tests, exploratory workflows and diagnostics that make delivery faster and more reliable.

- Improve observability. Develop explicit visualisations, monitoring and alerts that make model behaviour easier to understand and investigate.

- Partner with Data Science. Contribute to validation, feature engineering, feature extraction and methodology improvements based on what you learn from real customer deployments.

- Partner with Engineering. Improve data access, workflow automation, monitoring and the production reliability of the systems our models depend on.

- Strengthen how the team works. Contribute to peer review, documentation, shared standards and quality gates so knowledge and quality do not depend on one person.

- Communicate clearly. Make findings, uncertainty,



risks and trade-offs understandable to technical and non-technical colleagues.

What we’re looking for

You’ve worked in data science, applied statistics, analytics, ML operations or a related role and are comfortable working with real-world data and recurring analytical workflows. You can move between detailed investigation and the wider customer or business context, and you care about making sophisticated analysis reliable in practice.

Beyond experience, we care about how you operate:

- You have strong analytical foundations. You understand statistical uncertainty, hypothesis testing, Bayesian reasoning and common machine-learning approaches, and you know how to challenge your own conclusions.

- You’re fluent in Python and SQL. You can explore, transform and validate complex datasets and turn recurring work into maintainable tooling.

- You write code like an engineer. Your analysis doesn’t end in a notebook: you version-control your work, write tests, take review seriously, and produce code that colleagues can maintain and build on.

- You’re rigorous about quality. You validate results honestly, pay attention to detail and treat peer review and testing as essential rather than optional.

- You’re a systematic diagnostician. You can isolate the source of an issue and distinguish between a data problem, configuration error, engineering failure or modelling limitation.

- You’re comfortable with messy data. You can find structure, anomalies and useful signals in incomplete or inconsistent customer datasets.

- You understand production ML

📌 Production ML Scientist: Model Quality (Melbourne)
🏢 Mutinex
📍 Melbourne