I build Bayesian statistical methods and tools for early-phase oncology trials — from dose-finding designs to PK/PD-informed dose optimisation to cure-rate survival models.
Most of my work is hands-on: R and JAGS/NIMBLE implementations, simulation studies, Shiny apps, and manuscripts translating theory into methods trial teams can actually use.
- Bayesian dose-finding & dose-optimisation — seamless Phase I/II frameworks combining PK/PD modelling with toxicity/efficacy dose-finding
- Multi-agent combination designs — comparing designs (BOIN12, EffTox, COMIC, uTPI-Comb) for two-agent optimal biological dose identification
- Cure-rate & survival modelling — mixture cure models for long-term outcome trials with non-susceptible fractions
- Rare-disease trial methodology — restricted mean duration of response under non-proportional hazards, small-sample group-sequential designs
- JAGS ↔ NIMBLE model porting, prior elicitation from published trial data, and simulation-based operating characteristics
- PREDOSE — seamless Phase I/II Bayesian PK/PD dose-optimisation framework for oncology step-up dosing (manuscript targeting JASA)
- OBDC-Compare-App — Shiny app comparing four Bayesian dose-finding designs for two-agent combination trials
- PKComb-BOIN12 Simulator — Shiny app supporting PKComb-BOIN12: A Pharmacokinetics-Informed Bayesian Optimal Interval Design for Dose Optimization in Cancer Drug-Combination Trials (manuscript in preparation)
- Estimands-for-Cure-Rate-Model — estimand and causal framework for cardiovascular trials with non-susceptible fractions
- Associate Advisor – Statistics, Early Phase Oncology Biostatistics @ Eli Lilly and Company; trial statistician on early-phase oncology studies including an ADC program
- PhD in Statistics, Indian Statistical Institute, Kolkata — Bayesian joint modelling of multivariate longitudinal and time-to-event data
- Full member, PSI (Statisticians in the Pharmaceutical Industry)
- Published in Journal of Applied Statistics, Lifetime Data Analysis, and Journal of Biopharmaceutical Statistics
- Previously a data scientist at RedBus and Intuit before moving into biostatistics
- Core tools: R, JAGS, NIMBLE