Data Scientist at Postbank (SOC) Ltd

Purpose of the Job

• The Data Scientist is accountable for independently designing, developing, validating and operationalising advanced analytics and machine-learning solutions for assigned business problem areas. The role translates complex business and regulatory questions into analytical approaches, makes interpretive technical decisions within approved governance frameworks, and produces auditable, production-ready analytical outputs that support operational, customer, financial, risk and regulatory decision-making across Postbank.

Job Responsibilities
Advanced Analytics Solution Ownership

• Own assigned advanced analytics workstreams from problem framing to deployment-ready output.
• Translate business problems into analytical hypotheses, data requirements, modelling approaches and delivery plans.
• Design and build statistical, machine-learning and optimisation models using appropriate techniques and documented assumptions.
• Determine suitable validation, back-testing and monitoring approaches for assigned models and analytical outputs.
• Convert model and analysis results into clear business recommendations, implementation options and risk considerations for stakeholders.

Data Engineering, Quality and Reproducibility

• Acquire, integrate, profile and transform structured and unstructured data from approved sources.
• Design repeatable data preparation, feature engineering and analytical pipelines with appropriate quality controls.
• Maintain version-controlled code, model artefacts, assumptions, metadata, outputs and documentation for auditability and knowledge retention.
• Identify data-quality, lineage, completeness and integrity issues and recommend corrective actions to the Senior Specialist.

Model Risk, Governance and Compliance

• Apply Postbank data-governance, model-risk, POPIA, FSCA and internal information-security requirements.
• Conduct reasonableness checks, bias/variance assessments, explainability reviews and performance monitoring for assigned analytical outputs.
• Document model limitations, residual risks and recommended monitoring controls before implementation or handover.
• Support internal reviews, audit requests and regulatory evidence packs relating to analytical models and outputs.

Stakeholder Advisory and Delivery

• Engage business, operations, BI, automation, data and technology stakeholders to clarify requirements and implementation constraints.
• Present analytical findings to technical and non-technical audiences in a clear, evidence-based manner.
• Provide specialist input into analytics standards, reusable methods, reusable code libraries and continuous-improvement initiatives.
• Guide junior analysts or project contributors on technical methods where required, without formal people-management accountability.

Qualifications and Experience

• Bachelor’s degree in Mathematics, Statistics, Applied Mathematics, Data Science, Engineering, Economics, or a related quantitative field (NQF Level 7).
• Honours or Masters degree in a quantitative, data science, analytics or related field (NQF Level 8/9 (Ideal)
• 5+ Years Relevant data science, machine learning and algorithm-related experience at a specialist level
• Experience in the Banking Industry (Advantageous)

Skills and Attributes

• Uncompromising integrity, honesty, and professional ethics.
• Respectful and professional conduct at all times.
• Strong personal accountability and ownership.
• A commitment to excellence in every deliverable.
• High levels of resilience and perseverance when facing obstacles.
• Self-motivation and the ability to operate independently.
• Continuous learning and intellectual curiosity.
• Loyalty to team objectives and organisational success.

Phase Target, Skills & Algorithms, Tech Stack Mapping
Core Data &BI Engineering

• Database optimization, automated ETL pipelines, stored procedures. SQL Server (SSMS, SSIS), Power Query
• Advanced data modeling, context filtering, row-level security, cloud deployment. Power BI (DAX), Tableau
• Statistical profiling, macro migration, legacy predictive validation. SAS

Predictive AI & Automation

• Binary classification, feature engineering for upsell/crosssell propensity modeling. Python (scikit-learn, XGBoost), R, SQL Server
• Regression trees, ensemble learning, automated lead generation scoring engines. Python, SQL Server, Power Query Survival analysis, lifetime value (LTV) estimation, NLP for service ticket intent. Python (lifelines, HuggingFace), R, Power BI

Advanced Optimization & Geo

• Location allocation algorithms, spatial clustering, spatial regression. ArcGIS, Python (GeoPandas, PySAL)
• Constrained optimization, attribution modeling, marketing mix modeling (MMM). Python (SciPy.optimize), R
• Bridging deep tech with executive business decisions, setting architectural patterns. Full Ecosystem Transformation