Data Scientist

Be Part Of The World’s Largest Logistics Company

Deutsche Post DHL Group is the world’s leading logistics and mail company.

We are one of the world’s largest employers, operating in over 220 countries and territories. We are Europe’s largest postal service, partner for eCommerce and pioneers in secure digital communication. We are number one in contract logistics and international express delivery, and a leader in the forwarding business.

Join us and you will be working for a global company that is focused on service, quality, and sustainability, and using the power of global trade to connect people and improve lives.

And not just for our customers, but for every member of our Group too.

At DHL Supply Chain, we are looking for a business–minded and technical...

To design, build, and operationalize machine learning and optimization solutions that deliver measurable improvements in planning, execution, and service performance across warehousing and logistics. Translate high-value use cases into production-ready models with clear ROI and strong user adoption.

• Frame analytical problems, define success metrics, and design robust experiments (A/B tests or quasi experimental designs).
• Build data pipelines and validated training/evaluation datasets and establish and maintain feature stores as needed.
• Train, evaluate, and iterate on models (including tree based methods, time series, clustering, and deep learning when appropriate).
• Implement Machine Learning Operations practices, including experiment tracking, model registry, CI/CD, and monitoring for drift and performance degradation; automate retraining where feasible.
• Collaborate with Data Engineering to deliver reliable batch and API integrations into BI platforms and operational systems.
• Ensure Responsible AI practices: privacy by design (POPIA), bias and impact assessments, and comprehensive model documentation (e.g., model cards).
• Communicate insights clearly through data storytelling and support stakeholders via dashboards and intuitive UX integrations (e.g., Power BI).
• Comply with Information Security and POPIA requirements; enforce privacy by design throughout the machine learning lifecycle.
• Maintain model cards, data lineage, validation reports, and change logs to ensure audit readiness.

• Deliver at least one complex, one medium, and two simple projects per cycle with successful implementation outcomes.
• Achieve measurable improvements in targeted KPIs for each implemented solution.
• Realize cost reductions for each completed project, aligned with business case expectations.
• Maintain time to production of ≤ 12 weeks for priority models, with ≥ 80% adoption by intended users.
• Ensure model reliability of ≥ 99% uptime, with documented drift detection processes and defined retraining SLAs.

• Bachelor’s degree or equivalent experience (Statistics, Computer Science, Engineering, or related).
• Azure Machine Learning, Databricks, Machine Learning Operations; CI/CD with Azure Dev Ops (ADO) or Git Hub; testing (pytest).
• Strong Python skills (pandas, Num Py, scikit-learn), SQL; experience at scale with time-series and tabular Machine Learning.

• Minimum 2 years in applied data science; 3–7+ years preferred.
• Optimization (OR-Tools/Pyomo); forecasting frameworks; SHAP/explainability.
• Experience embedding model outputs into WMS/TMS/ERP workflows and Power BI.
• Data cleansing and processing; databases/data warehousing and integration; cloud computing (Azure preferred).
• Predictive modelling and machine learning; statistics and mathematics; workflow automation.
• Business decision-making, stakeholder management, consulting, change management, presentation and storytelling.
• Data visualization (Power BI); data protection and privacy regulations; technology trends; project management.

• Stakeholder management.
• Identifying and analysing information to understand project requirements, key data, and decision drivers.
• Applying technical skills to create and model scenarios for current and future operational solutions.
• Developing value based solutions with a focus on…