Swisscontact is an independent non-profit development organisation established in 1959 by representatives of the Swiss private sector and civil society. Our work focuses on private sector-led, sustainable economic development and improved quality of life for all in developing and emerging countries. A key focus is to strengthen the skills of individuals and foster the competitiveness of businesses.
Tasks:
- End-to-End Data Products: Architect and build Fabric-native data solutions—OneLake-backed Lakehouses and Datamarts, medallion-layer models, and governed semantic models - for Finance, CRM, HRM, and M&E domains.
- Analytics & BI: Design high-quality Power BI datasets and reports (DAX/Tabular) with row-level security, certified datasets, and metrics for executive and operational decision-making.
- ML/AI Delivery: Develop, train, and deploy Python models (classification, regression, forecasting, NLP, recommender systems). Implement feature stores, model versioning, and automated retraining.
- Pipelines & Orchestration: Build reliable ingestion and transformation with Fabric Data Factory/Synapse (Spark/SQL). Enforce CI/CD, testing, lineage, and monitoring.
- Data Quality & Governance: Define data contracts, validation rules, and documentation; champion privacy, compliance, and responsible AI.
- Collaboration & Enablement: Partner closely with Finance, Business Development, HR, and M&E teams to translate requirements into technical designs. Coach M&E colleagues, and data consumers across the organization; Promote best practices.
Requirements:
- Education: Bachelor’s or higher in Data Science, Computer Science, Statistics, Engineering, or a related field.
- Experience: 4–5+ years delivering analytics and ML solutions
- Platform Expertise: Microsoft Fabric: OneLake, Lakehouse (Delta), Data Engineering (Spark), Data Factory pipelines, Synapse SQL. Power BI: Data modeling, DAX, performance optimization, governance (certified datasets, RLS). Python for ML/AI: pandas, numpy, scikit-learn, PySpark; solid understanding of model design, evaluation, and MLOps.
