AIML Engineer, Agentic AI COA Accelerator

IQVIA is a leading global provider of clinical research services, commercial insights, and healthcare intelligence to the life sciences and healthcare industries. Their Clinical Outcomes Assessments (COAs) organisation focuses on incorporating the patient voice into medication and intervention development. The COA Accelerator ecosystem is expanding AI-enabled capabilities to support evidence-backed COA strategy development and expert decision support.

• Degree in computer science, machine learning, artificial intelligence, data science, computational linguistics, biomedical informatics, bioinformatics, engineering, or related field.
• Experience building AI, NLP, LLM-powered, or machine learning applications in production.
• Practical experience with retrieval-augmented generation, embeddings, vector databases, semantic search, prompt engineering, model evaluation, and LLM orchestration.
• Strong Python skills and familiarity with modern AI/ML frameworks, APIs, testing practices, version control, and deployment.
• Experience working with unstructured scientific, clinical, regulatory, healthcare, or research data.
• Ability to translate expert reasoning into technical implementation with nuanced, evidence-based domain logic.
• Strong understanding of AI output quality risks such as hallucination, overconfidence, weak grounding, inconsistent reasoning, and unsupported recommendations.
• Strong analytical judgement to identify incomplete or misleading AI-generated outputs requiring human review.
• Ability to collaborate with technical and non-technical stakeholders.
• Excellent documentation skills explaining AI system behaviour, assumptions, limitations, validation, and governance.

Additional Requirements:

• Preferred experience in life sciences, clinical research, regulatory strategy, COAs, patient-reported outcomes, medical evidence synthesis, or clinical decision support.
• Experience with AI orchestration frameworks like LangChain, LlamaIndex, Haystack, or Semantic Kernel.
• Experience with vector databases/search technologies such as Azure AI Search, Pinecone, Weaviate, Milvus, Qdrant, OpenSearch, or Elasticsearch.
• Familiarity with model evaluation tools, LLM observability platforms, automated evaluation frameworks, benchmarking, and continuous improvement.
• Desirable experience with fine-tuning or adapting models using LoRA, QLoRA, or instruction tuning.
• Experience with cloud environments like Azure, AWS, or GCP, preferably in regulated healthcare or life sciences settings.
• Understanding of GDPR, data privacy, secure AI deployment, proprietary data handling, access controls, auditability, and governance.
• Ability to work independently in remote or hybrid environments while collaborating globally.
• Significant experience with AI tools like Cursor, Claude Code, or Codex.
• Fluency in English.

• Design, develop, and optimise AI capabilities for COA strategy generation, evidence synthesis, and recommendation development.
• Implement AI reasoning, retrieval-augmented generation, model orchestration, prompt architecture, and domain adaptation.
• Build AI workflows interpreting therapy area, indication, study phase, endpoint objectives, regulatory context, and trial design.
• Develop AI capabilities generating structured COA strategy recommendations with rationale and supporting evidence.
• Design and optimise retrieval pipelines across structured and unstructured sources including COA libraries, scientific literature, regulatory labels, and clinical trial records.
• Implement semantic search, hybrid search, metadata filtering, reranking, source attribution, and citation-supporting workflows.
• Evaluate usage of foundation models, fine-tuned models, embeddings, rerankers, rules, or hybrid approaches.
• Develop mechanisms to discern evidence strength and areas requiring expert judgement.
• Create and maintain model evaluation frameworks addressing factuality, retrieval relevance, citation accuracy, and hallucination risk.
• Partner with scientists, product managers, data and software engineers, security, legal, and commercial teams.
• Support demos, prototypes, pilots, and client-facing proof-of-concept work with clear AI functionality explanations.
• Document model behaviour, assumptions, limitations, evaluation results, and decision logic.
• Contribute to AI governance practices for responsible AI use in clinical research and regulated decision-support contexts.

IQVIA maintains a zero tolerance policy for candidate fraud; all submitted information must be truthful and complete. False statements or omissions will result in application disqualification or employment termination. IQVIA is committed to diversity, inclusion, equality between women and men, and responsible AI governance.