Lead Data Scientist – (Azure Databricks)

Role: Lead Data Scientist (Azure Databricks) Location: Tampa / Austin Type: Contract/FTE Role Summary • As Lead Data Scientist, you will spearhead the end-to-end development of sales forecasting and demand sensing models for CPG portfolios on Databricks (Azure). You will work closely with commercial, supply chain, and engineering teams to build ML solutions that improve forecast accuracy, reduce inventory waste, and support revenue growth. You bring deep ML expertise, strong Python engineering skills, and a nuanced understanding of CPG market dynamics and you are comfortable translating complex model outputs into clear business recommendations. Job Description • 1. Lead end-to-end sales forecasting model development from data sourcing and feature engineering through model training, validation, and productionisation on Databricks (Azure). • 2. Design and maintain forecasting pipelines at SKU, category, and regional hierarchy levels incorporating POS data, promotional calendars, seasonality indices, and external signals (macroeconomic, weather). • 3. Apply CPG domain knowledge to model promotional uplift, new product introduction curves, product cannibalization, and retailer sell-in/sell-out dynamics into ML features and targets. • 4. Operationalise ML models using MLflow on Databricks manage the model registry, version control experiments, automate retraining schedules, and configure drift monitoring alerts. • 5. Collaborate with commercial and supply chain teams to translate forecast outputs into inventory recommendations, production planning inputs, and revenue growth strategies. • 6. Define and enforce data science best practices modelling standards, experiment documentation, code review guidelines, and reproducibility requirements across the team. • 7. Mentor junior data scientists conduct code reviews, lead knowledge-sharing sessions, support career development, and build a high-performance data science culture. • 8. Communicate model insights and forecast accuracy to senior stakeholders through dashboards, executive briefings, and written reports making complex model behaviour accessible to business audiences. • 9. Drive continuous model improvement benchmark new algorithms, evaluate AutoML approaches, and run controlled experiments to improve MAPE, bias, and coverage metrics. • 10. Partner with data and platform engineers to ensure feature pipelines on Azure Data Lake / Delta Lake are reliable, scalable, and aligned with model refresh cadence requirements. Primary (Must have skills) • • 3+ years of experience in Databricks in production • 5+ years of experience in Python pandas, PySpark, scikit-learn • 5+ years of experience with Azure ML or Azure ecosystem • 3+ years of experience in MLflow or equivalent experiment tracking tool • 5+ years of experience in Supervised, unspervised machine learning algorithms, forecasting and inventory optimization • 5+ yeras of experience in deep learning algorithms applying to solve forecasting, regression and classification problems • 3+ years of experience in buidling ML models in CPG industry Secondary Skills (Good To Have) • Statistical Analysis & Experimentation • A/B testing, causal inference, and hypothesis testing to measure the business impact of model improvements and pricing interventions. • SQL & Data Engineering Fundamentals • Advanced SQL on Delta Lake / Azure Synapse; ability to build lightweight feature pipelines without full data engineering support. • MLOps & CI/CD for ML • MLflow, GitHub Actions, or Azure DevOps pipelines to automate model retraining, evaluation gates, and deployment to Databricks Model Serving. • Data Visualisation & Storytelling • Power BI, Plotly, or Streamlit dashboards to communicate forecast accuracy and business KPIs to non-technical stakeholders. • Promotional & Trade Analytics • Modelling promotional uplift, baseline vs incremental volume splits, and trade spend ROI key for CPG forecast decomposition • Team Leadership & Mentoring • Guide junior data scientists, run code reviews, define modelling standards, and represent the data science function in cross-functional forums.