Desi R. Ivanova

Desi R. Ivanova is a Florence Nightingale Bicentennial Fellow in the Department of Statistics at University of Oxford. Her research focuses on robust evaluation of large language models and Bayesian experimental design. Ivanova holds a DPhil in Statistics (StatML CDT) from Oxford and previously worked in quantitative finance at Goldman Sachs.

Research Interests: Ivanova develops information-theoretic methods for contextual optimization and causal dataset acquisition. Current work addresses statistical rigor in AI evaluation, semantic calibration, and privacy-preserving data integration.

Publications: Her contributions introduce novel frameworks for adaptive experimental design (DAD, Step-DAD) and critically examine evaluation practices for foundation models.

Additional: Organizes ELLIS RobustML workshops and contributes to uncertainty quantification methods. Research integrates Bayesian principles with modern AI challenges.