Real AI applied to Real Healthcare

Faris' research career has focused on utilizing big data and machine learning techniques to both maximize clinical support and understand the biological underpinnings of complex disease. Developing tools for tasks like sample size estimation for machine learning and time series analysis, Faris has focused on improving accessibility to machine learning algorithms for practitioners across the spectrum. By bridging clinicians to machine learning, Faris has focused on generating novel insights by combining both knowledge of these experienced practitioners and an understanding of what machine learning can do. On the flip side, Faris has also focused on understanding the causal nature of genetics is disease, working with massive datasets such as the UK Biobank to reveal how complex disease expresses itself variably in the general population.

View Research Get in Touch

About Faris Gulamali

Faris Gulamali is an AI researcher and Interventional Radiology Resident with publications in NPJ DM, Communications Medicine, Science, ICLR, JMLR, JMIR.

After completing a undergraduate studies in Chemistry and Statistics at Williams College and medical school at the Icahn School of Medicine at Mount Sinai Hospital, Faris is moving to Duke for an Integrated Residency in Interventional Radiology.

Faris has published 15 articles with 3 patents, developing algorithms applied to neuro-critical care, nephrology and cardiology.

Areas of Expertise

Time Series for Healthcare

Developing novel state-of-the-art algorithms for time series clustering and classification

Federated Learning

Training algorithms across multiple hospital systems without violating patient privacy.

Non-invasive monitoring

Creation of AI systems that assist physicians in detecting and characterizing abnormalities in real-time settings.

Procedural AI Assistance

Pioneering AI systems that provide real-time guidance during interventional procedures, enhancing precision and reducing complications.

Sample Size Estimation

Advanced techniques for sample size estimation in the healthcare space of complex AI algorithms.

Data-Centric AI

Research focused on debiasing AI decision-making via data-centric approaches.

Current Research

AI-Enhanced Early Detection of Intracranial Hypertension

Developing deep learning algorithms that can identify subtle biomarkers of qntracranial hypertension quicker than conventional methods, potentially improving treatment outcomes.

Real-time AI Guidance for Complex Vascular Interventions

Creating an AI system that provides dynamic guidance during complex vascular procedures, predicting optimal tool paths and identifying potential complications before they occur.

Debiasing AI algorithms via data-centric approaches

Detecting, characterizing and mitigating a breadth of differentiable and non-differntiable algorithms via proxies for learnability.

Derivation, external and clinical validation of a deep learning approach for detecting intracranial hypertension

Gulamali et al

npj Digital Medicine 7.1 (2024): 233
Online Unsupervised Representation Learning of Waveforms in the Intensive Care Unit via a novel cooperative framework: Spatially Resolved Temporal Networks (SpaRTEn).

Gulamali, F. F., Sawant, A. S., Hofer, I., … Nadkarni, G. N.

Journal of Machine Learning Research, Machine Learning in Healthcare Conference, 1-22, 219.
The genetic architecture and evolution of the human skeletal form

Kun, Eucharist, Emily M. Javan, Olivia Smith, Faris Gulamali, Javier de la Fuente, Brianna I. Flynn, Kushal Vajrala et al.

Science 381, no. 6655 (2023): eadf8009. (front cover)
Autoencoders for sample size estimation for fully connected neural network classifiers

Gulamali, F. F., Sawant, A. S., Kovatch, P., Glicksberg, B., Charney, A., Nadkarni, G. N., & Oermann, E

npj Digital Medicine, 5(1), 180.
Federated Learning in Risk Prediction: A Primer and Application to COVID-19 Associated Acute kidney Injury

Faris F Gulamali. Girish N. Nadkarni

Nephron. DOI: 10.1159/000525645.