Data Engineer

Summary

Theoris Services is assisting a client with hiring a Data Engineer for a Clinical Trial Foundations engagement focused on preparing domain datasets for AI use. The role builds and maintains data pipelines and models, embeds data quality and lineage, documents metadata, captures business questions, establishes ground truth, and evaluates AI model performance against clinical datasets.

Responsibilities

• Build and maintain ETL/ELT pipelines and support infrastructure for Clinical Trial Foundations datasets
• Design and implement data models that support availability, performance, and reuse for AI querying
• Embed data quality checks, performance considerations, and lineage so datasets are reliable inputs for downstream AI use
• Audit existing data products to document schema, field definitions, data types, and business meaning
• Identify gaps where schema or metadata is missing, inconsistent, or undocumented — a primary blocker to AI readiness
• Tag datasets with sensitivity and classification levels (PII, GxP-regulated, public, and related categories)
• Work directly with business stakeholders to capture the specific questions they want the data to answer
• Translate ambiguous business asks into structured, testable questions a dataset should be able to support
• Maintain a living inventory that maps datasets to the business questions they support or should support
• Establish ground truth answers and values for representative business questions per dataset
• Design evaluation sets (question and correct-answer pairs) to test whether AI systems querying the data return accurate results
• Identify and document known data quality issues, edge cases, and limitations that could cause inaccurate AI responses
• Use Claude (or similar models) to generate responses and insights against datasets, then evaluate those responses against ground truth
• Score and categorize failure modes (hallucination, stale data, misinterpreted schema, wrong aggregation logic, and related issues)
• Iterate on metadata and schema documentation based on where model responses fail, and feed those fixes back into dataset AI-readiness
• Build repeatable evaluation rubrics and scorecards that can scale across the domain’s datasets
• Document metadata, ground truth sets, and evaluation results centrally so other teams do not repeat the work

Skills

• Direct experience in Clinical Trial Foundations or related clinical development domains, including protocols, clinical trial data, regulatory submissions, or medical affairs
• Working knowledge of clinical and medical data standards (for example CDISC, SDTM, MedDRA) and GxP-regulated data handling
• Familiarity with pharmacovigilance, safety, or regulatory reporting data flows
• Experience building ETL/ELT pipelines, data models, and data quality controls, with designs oriented toward data availability for AI
• Experience building AI/LLM evaluation frameworks (for example with Claude or similar models) against domain-specific datasets
• Ability to translate clinical and medical stakeholder questions into structured, testable data requirements
• Strong documentation habits for schema, metadata, evaluation sets, and reusable scorecards
• Clinical, medical, pharmacovigilance, or regulatory data background is strongly preferred over general data engineering experience alone

Qualifications

Must Haves

• Direct experience in Clinical Trial Foundations or related clinical development domains, including protocols, clinical trial data, regulatory submissions, or medical affairs
• Working knowledge of clinical and medical data standards (for example CDISC, SDTM, MedDRA) and GxP-regulated data handling
• Familiarity with pharmacovigilance, safety, or regulatory reporting data flows
• Experience building ETL/ELT pipelines, data models, and data quality controls, with designs oriented toward data availability for AI
• Experience building AI/LLM evaluation frameworks (for example with Claude or similar models) against domain-specific datasets
• Ability to translate clinical and medical stakeholder questions into structured, testable data requirements
• Strong documentation habits for schema, metadata, evaluation sets, and reusable scorecards

Nice to Haves

• Clinical, medical, pharmacovigilance, or regulatory data background is strongly preferred over general data engineering experience alone

Benefits

• Remote work arrangement
• Robust Health Insurance
• 401(k) plan
• PTO accrual
• Paid holidays
• Excellent cash-based referral program