Machine Learning Developer

About this role:

We are looking for a mid-level Data Scientist (4+ years of experience) to help build and maintain forecasting models. You will work with time series methods such as ARIMA and gradient-boosted models such as XGBoost, developing solutions natively within Microsoft Fabric. This role suits someone who enjoys translating real-world financial data into reliable, production-ready forecasts.

What You Will Do:

• Design, build, and validate time series forecasting models (ARIMA, exponential smoothing) and machine learning models (XGBoost) for various business forecasting processes.
• Utilize data pipelines within Microsoft Fabric to ingest, clean, and transform structured and time-indexed data.
• Evaluate model performance using appropriate statistical and ML metrics (MAPE, backtesting) and iterate on feature engineering and model selection.
• Suggest and implement recommendations to enhance model outcomes/additional models as data structures and business needs evolve.
• Deploy and monitor models in production, ensuring forecasts refresh reliably on schedule within Fabric pipelines.
• Partner with the ML team and other stakeholders to translate forecasting requirements into modeling approaches and communicate results and model limitations clearly to non-technical audiences.
• Document modeling assumptions, and validation results to support governance and audit requirements.
• Stay current with developments in time series and ML forecasting methods and recommend improvements to existing modeling practices.
• Support the wider ML team with utilizing the model outputs in business-oriented user interfaces.

What You'll Bring:

• 4+ years of experience in a data science, quantitative analytics, or applied statistics role, ideally with exposure to financial.
• Hands-on experience building and tuning ARIMA (or similar classical time series) models and XGBoost (or comparable gradient-boosting) models.
• Working knowledge of Microsoft Fabric.
• Proficiency in Python (pandas, stats models, scikit-learn, xgboost) and/or PySpark, solid SQL skills.
• Understanding of core statistical concepts: stationarity, seasonality, autocorrelation, cross-validation for time series, and bias-variance trade-offs.
• Strong communication skills, with the ability to explain modeling choices and forecast uncertainty to non-technical stakeholders.
• Bachelor's or Master's degree in Data Science, Statistics, Mathematics, Computer Science, or a related quantitative field.

Nice to Have:

• Familiarity with MLOps practices (model versioning, CI/CD for ML, monitoring for drift) within Fabric or Azure ML.
• Exposure to actuarial, insurance, or macroeconomic forecasting contexts.