About the role
Responsibilities:
- Build and maintain production services and APIs that serve analytics and ML model outputs at scale across multi-country GCC deployments.
- Design and implement ETL/ELT pipelines feeding analytical and ML workloads including supply chain demand forecasting, claims adjudication signals, and population health indicators.
- Develop predictive and statistical models for health supply chain commodity flows, eClaims payer analytics, and population health use cases.
- Collaborate with product and domain teams to translate clinical and operational questions into well-scoped data problems with measurable success criteria.
- Own model monitoring, retraining pipelines, and data quality frameworks, not just initial deployment.
- Contribute to and extend the data lakehouse architecture built on Apache Iceberg, ClickHouse, and Parquet with Dagster orchestrating the pipeline.
- Participate in architecture decisions around the analytics data stack, including ClickHouse DR, lakeFS as Iceberg REST catalogue, and dbt-clickhouse transformations.
- Write clean, testable, production-grade code, not research notebooks passed to an engineering team.
Requirements:
- 6-9 years of total experience with a demonstrable split between data science and backend or data engineering, not purely one or the other.
- Strong Python (non-negotiable): Pandas, NumPy, scikit-learn, SQLAlchemy, and FastAPI or equivalent async frameworks.
- Solid SQL and hands-on experience with columnar or analytical databases; ClickHouse strongly preferred; BigQuery, Snowflake, or DuckDB accepted.
- Production model deployment experience: you have shipped models to live environments and maintained them, not just trained and handed off.
- Familiarity with Kafka or event-driven architectures is a meaningful plus.
- Experience with dbt or equivalent transformation tooling and an understanding of medallion or lakehouse architecture patterns.
- Experience in health data - FHIR, claims, supply chain, or logistics is a strong differentiator but not a hard requirement.
- Comfort working in a B2G context where data sovereignty, auditability, and schema compliance are first-class concerns
Nice to Have:
- Exposure to GS1 or EPCIS data formats for pharmaceutical or medical device traceability.
- Experience with Dagster, Prefect, or Airflow for pipeline orchestration.
- Familiarity with lakeFS, Apache Iceberg, or OpenTable format ecosystems.