Perth folks — we're hiring.
Applied Data Scientist / ML Engineer (Math & Statistics + General ML)
Location: Perth, WA
Work arrangement: On-site
Employment Type: Full-time
Experience: 5+ years in Applied ML/Data Science (or equivalent)
Eligibility: AU working rights required
What you'll do:
* Build decision-focused ML models used in our products (scoring / ranking / prediction).
* Work across several problem types, including:
geospatial scoring + confidence/risk maps
valuation modelling + "evidence" via comparable examples
explainable signals from time-dependent data (trend / risk / anomalies)
search/retrieval relevance ranking + dedup + exportable outputs
* Own modelling correctness and evaluation discipline: prevent overfitting, control leakage, robust validation, stability checks, error analysis.
* Own confidence/uncertainty in outputs: calibration, prediction intervals where relevant, clear limits of validity, "unknown/abstain" rules when needed.
* Deliver explainability that matches model behaviour: "why this result", assumptions, and failure modes (not black-box hope).
* Produce integration-ready ML artifacts (code + evaluation + documentation) for engineering to ship.
Typical model families you'll use (examples):
* Tabular ML: regularised regression / GAM, Gradient Boosting (CatBoost/LightGBM/XGBoost)
* Ranking & retrieval: hybrid retrieval (BM25 + embeddings), reranking / learning-to-rank
* Uncertainty tooling: calibration, quantile/interval methods, conformal-style approaches
* Deep Learning
Must-have:
* Strong mathematics + statistics (first principles): probability & inference, hypothesis testing (or Bayesian equivalents), bias/confounding, uncertainty & calibration, optimisation intuition. You understand overfitting vs generalisation and can prove it via correct evaluation.
* Strong general ML fundamentals: supervised learning, feature engineering, model selection, evaluation, robustness.
* Deep Learning hands-on: PyTorch or TensorFlow (training + inference basics).
Python suitable for a real codebase (clean, testable, PR-ready).
* Fast learner: can ramp quickly into new domains and collaborate with domain experts.
* Clear communication: can explain trade-offs and uncertainty to non-specialists.
Good to have:
* Time series / signals experience (time-aware validation, non-stationarity, anomalies).
* Any of: geospatial/spatial data, valuation/pricing, ranking/search/retrieval systems.
* Strong SQL.
You must have working rights in AU.
To apply, please send your CV to