Our commitment to science

Our team, comprised of esteemed scientists and technologists, is dedicated to pioneering work that stands at the forefront of healthcare and AI. Our research is constantly published in peer-reviewed journals, underscoring our commitment to contributing valuable insights to the scientific community. We take pride in being respected members of the academic world, continuously pushing the boundaries to further scientific efforts and improve patient care through our AI-enabled platform

Our research focus

At the heart of our efforts is the integration of artificial intelligence (AI) into clinical settings, with a particular emphasis on managing and improving outcomes for patients with chronic illnesses. We are committed to developing AI algorithms that not only analyze complex health data but also provide actionable insights for early intervention. Our work aims to shift the paradigm from reactive to proactive management in chronic disease care, focusing on prediction, prevention, and personalized treatment plans to fundamentally transform patient outcomes.

Recent work

Differentially Private Time Series EHR Generation (DP-TimeGAN)

In collaboration with the University of Oxford, we published a novel, differentially private model for generating realistic synthetic longitudinal health records to safely unlock restricted medical data for research.

Results

Generates realistic synthetic EHRs with mathematical privacy guarantees.
Validation showed model output maintained the critical patterns clinicians rely on and clinical reviewers could not distinguish synthetic data from real-world records.
Published at ML4H 2025 and NeurIPS 2025 (TS4H Workshop).
Major step toward building clinical AI systems that are private by design and ready for real-world deployment and impact.

SIAM Mathematical Problems in Industry Workshop 2026

Vironix was a problem presenter at the SIAM MPI workshop conducted at Drexel University. We presented a problem on developing disease degradation models for chronic digestive conditions, such as Inflammatory Bowel Disease, Irritable Bowel Syndrome, and Gastroesophageal Reflux Disease.

Highlights

Created mechanistic models to predict disease load and symptom flares.
Developed ML models to feed patient-specific parameters into the mechanistic models.
Final report to be published Fall 2026.

Oxford Collaboration & Gastroenterology Expansion

Partnered with the University of Oxford to develop our first chronic digestive disease models, expanding our clinical scope into gastroenterology alongside our foundational mathematics research.

Highlights

Built our first predictive models for flare-up focused specifically on chronic digestive diseases.
Work complements modeling work conducted at the SIAM Math Problems in Industry Workshop.
Research will be published as a MMSc thesis Fall 2026.

Synthetic Electronic Health Record Generation (ScoEHR)

In collaboration with Oxford University, ScoEHR was developed as a new approach to synthetic data generation in medical context to overcome issues with past methods using GANs.

Results

First diffusion based synthetic EHR generation model.
The state-of-the-art deep learning model for EHR generation.
Published at MLHC 2023 (acceptance rate of 33.6%).

Publications

Scientific Collaborations