Computer Systems Engineer AI
Description
About the role
Our client, a global biopharmaceutical company, is looking for an AI CSV / CSA Validation Analyst to join its GxP AI Agent portfolio. You will receive portfolio priorities and assignments from the client's R&D Data Operations – Global Data & AI organization, and work with AI Engineering, Business owners, CSV and Quality Assurance.
Role objective
Perform and coordinate risk-based CSV/CSA for custom GxP AI/GenAI applications, translating AI technical failure modes into defensible requirements, controls, testing, evidence, and release decisions. This is a hands-on validation analyst role.
Required prerequisites
• GxP CSV/CSA: Demonstrated delivery of regulatory risk assessments, GxP intended-use and electronic-record assessments, validation plans, requirements/traceability, IQ/OQ/PQ or equivalent protocols, deviations, evidence reviews, validation reports, and change controls; familiarity with applicable GAMP 5, FDA CSA, and 21 CFR Part 11 principles.
• GenAI architecture: Practical understanding of LLMs, RAG pipelines (document ingestion, chunking, embeddings, retrieval and grounding), AI agents, prompts, tool calls, orchestration and downstream integrations.
• Probabilistic system testing: Ability to differentiate deterministic requirements from nondeterministic outputs and develop repeatable, risk-based evaluation strategies with representative test sets, variability checks, negative tests and defensible pass/fail criteria.
• AI-specific risks: Ability to assess hallucinations, unsupported claims, omission, prompt injection, inappropriate agent actions, autonomy, human review/approval controls and impact on patient safety, GxP records and data integrity.
• AI output evaluation: Hands-on capability to design tests for factual correctness, completeness, relevance, consistency, citation/source accuracy, retrieval quality and measurable acceptance thresholds for clinical/regulatory use cases.
• Data integrity and traceability: Assess source quality, provenance, study-level access, transformations, version/cutoff control, audit trails and traceability from source material through retrieved context to generated outputs.
• Lifecycle/change controls: Evaluate impacts of model, prompt, retrieval-index, source-data, configuration and integration changes; define regression testing, approvals and monitoring appropriate to risk.
Responsibilities and deliverables
• Develop and coordinate AI system RRAs and validation strategies with Business, AI Engineering and independent QA review; identify risk controls, intended-use boundaries and residual risks.
• Prepare and coordinate requirements, test protocols, evaluation scenarios, traceability, execution evidence, deviation resolution, validation summary and change-control packages.
• Plan and review system, interface and end-to-end testing, including downstream agent behavior after source, model or retrieval changes; ensure audit-ready documentation.
• Facilitate technical validation discussions, track deliverables and escalate material risks or missing evidence; work within client-approved SOPs, templates and systems.
Skills
Sign in to read and join the discussion.