About the role
AI summarisedThis internship involves building an AI-enabled Bayesian decision intelligence platform for oncology clinical trials. The role combines Bayesian statistics, machine learning, and large language models to model efficacy, safety, priors, external controls, dose-finding, and pharmacokinetics, while generating insights and communicating results to technical and non-technical stakeholders.
BiotechOnsite
Key Responsibilities
- Develop Bayesian models for primary and key secondary endpoints (e.g., response, PFS/OS, safety event rates), including credible intervals and posterior probabilities as decision summaries
- Implement ML models to predict response and safety based on baseline features/biomarkers collected during patient screening process, which may inform inclusion/exclusion criteria for enrolment
- Construct and justify priors informed by historical trials and real-world evidence; quantify prior influence (e.g., effective sample size) and run sensitivity analyses across alternative priors
- Implement borrowing methods across similar diseases or disease subtypes, and borrowing information between subgroups of a patient population
- Build ML-based patient similarity or propensity matching scores and confounding adjustment pipelines to support use of external controls
- Develop Bayesian dose-finding models (e.g., model-based or model-assisted approaches) and optimize dose selection balancing efficacy/toxicity
- Implement methods for population PK and PK-ADA covariate modelling, exposure-adjusted safety and efficacy modelling
- Create an internal 'Clinical Insights Copilot' to draft reports such as weekly signal summaries or executive-ready briefs from modelling results
- Produce stakeholder-ready slide decks to clearly explain: (1) the clinical question, (2) the modeling approach with intuitive visuals, (3) assumptions + limitations, (4) key results with uncertainty, and (5) recommended actions to both technical and non-technical audiences
Requirements
- Currently enrolled in a Bachelor’s or Master’s Program in a related field
- Possess strong R skills for clinical statistics and reporting (tidyverse, stat libraries)
- Familiarity with Bayesian modelling concepts: priors, posterior inference, operating characteristics via simulation, sensitivity analyses, documentation
- ML experience for clinical outcome prediction (tree-based models, xgboost, survival ML, calibration/validation, interpretability)
- Experience with modern AI tooling (LLM/RAG frameworks) and responsible use for patient and sensitive data (human-in-the-loop, auditability)
- Experience with LLM/AI frameworks for the development of novel analytics
- Understanding of data privacy and security best practices for patient data
- Interest in oncology drug development and clinical/biostatistical approaches
- Excellent communication skills to present complex concepts in a clear manner
- Ability to commit to a 20-week internship period