Data Engineer

For our client, one of the Global Pharmaceutical Company, we realize a recruitment process for Data Engineer role.

• At least 5 years of experience as Data Engineer.
• Strong data engineering skills: Databricks, Spark, Delta Lake, SQL, ETL design and orchestration.
• Familiarity with clinical trial concepts (inclusion/exclusion criteria, endpoints, demographics) and biomedical terminologies.
• Practical experience with data modeling and working with end users to define requirements.
• Experience with CI/CD, testing frameworks, and monitoring for data pipelines and ML models.
• Experience with NLP for information extraction from scientific text (publications, registries).
• Fluency in English both written and spoken.

Nice to have:

• Experience with pharmaceutical sector or clinical research data environments.

• Combine clinical data expertise with strong data engineering and technical skills to generate well documented pipelines from source to curated data sets in common data models like CDISC SDTM.
• Collaborate closely with clinical SMEs, data scientists, infrastructure, and other skilled data engineers.
• Include external benchmarking data as a FounData product and help automate extraction and harmonisation of competitor clinical trial data from public registries and publications into structured, analysis ready formats.
• Productionise and monitor pipelines and models; collaborate on CI/CD, testing, and user feedback.
• Implement ETL patterns (medallion architecture), ensuring data provenance, validation, and versioning.
• Take part of continuous improvement and validation of existing pipelines.
• Ensure clinical concepts are correctly represented and harmonized across data models (CDISC SDTM/ADaM, OMOP, HL7); contribute to mapping and transformation logic.
• Develop NLP models for entity and relation extraction (e.g., inclusion/exclusion criteria, demographics, endpoints, study design).
• Build automated pipelines to ingest registry and publication data and convert to tabular, queryable datasets.
• Co design the benchmarking data model with end users and map extracted information to standardized terminologies.
• Integrate human in the loop review, confidence scoring, and vocabulary/units normalization.