Elly Kipkogei

image_profile.jpeg

I am a third year PhD student in the Department of Biostatistics at Columbia University working under the mentorship of Prof Bin Cheng and Prof Min Qian .

Before joining Columbia, I worked at AstraZeneca (AZ) BioPharmaceutical Company as a Senior Data Scientist. At AZ, I researched explainable AI (xAI) methods and their applicability to clinical factors, molecular variables, and clinical outcomes as opposed to text or image data. Together with my collaborators, our objective was to exploit the ability of deep learning models to ingest datasets from multiple modalities, predict survival outcomes and identify prognostic variables. For explainability part of our research, we looked at attention mechanism - a defining characteristic of transformer-based deep learning models. I also worked on methods for identifying predictive biomarkers in cancer treatments.

My research interests revolve around 3 main areas:

  • Explainable AI: Leveraging predictive power of deep learning and the use of attention-based transformer models in cancer drug development.

  • Clinical Trials: AI and clinical trials, design, implementation and evaluation of randomized trials within the hospital settings.

  • Personalized Medicine: Flexible methods for identifying predictive biomarkers in Oncology.

selected publications

  1. Pretrained transformers applied to clinical studies improve predictions of treatment efficacy and associated biomarkers
    Gustavo Arango-Argoty, Elly Kipkogei, Ross Stewart, Gerald J Sun, Arijit Patra, Ioannis Kagiampakis, and Etai Jacob
    Nature Communications, 2025
    Bib PDF
  2. AI-driven predictive biomarker discovery with contrastive learning to improve clinical trial outcomes
    Gustavo Arango-Argoty, Damian E Bikiel, Gerald J Sun, Elly Kipkogei, Kaitlin M Smith, Sebastian Carrasco Pro, Elizabeth Y Choe, and Etai Jacob
    Cancer Cell, 2025
    Bib PDF