DNA Data Storage
We research DNA as an archival storage medium, from encoding error-corrected data into DNA to synthesis, storage, and ML-based decoding. We also explore DNA nanostructures and biomolecular computing for in-storage processing.
Our work emphasizes scalable, robust coding schemes and efficient indexing and random access for retrieving data from large DNA pools. See full project website here.
Selected publications:
Selected publications:
AI and Data Management for Healthcare
We study data management and AI for healthcare data. A key focus is applying explainable AI and machine learning to assisted reproductive technology (ART), helping clinicians personalise IVF treatment.
Our work in this context spans follicle size optimisation at trigger, AI-driven drug dosing, and deep learning for embryo and oocyte quality assessment—leveraging large, multi-centre clinical datasets.
Selected publications:
Selected publications:
AI for Data Management
We develop learned index structures — replacing classical data structures such as B-trees with ML models trained on data distributions.
Our theoretical and systems work spans constant-time query guarantees, information-theoretic complexity measures, and practical streaming indexes for real-time database environments.
Building on this, we investigate model architectures and training regimes that adapt continuously to evolving data distributions while preserving strict latency bounds.
Selected publications:
Selected publications: