Yishai Shimoni

Overview

Yishai Shimoni

Pronouns

He/Him/His

Title

Manager and Senior Research Scientist, Causal Machine Learning for Healthcare

Location

IBM Research - Israel Givatayim, Israel

Bio

I am the manager of the Machine Learning for Healthcare and Life Sciences group at IBM Research - Israel and co-leading the Biomedical Foundation Model (BMFM) effort, with a focus on the BMFM-targets subtheme. In these roles, I am in charge of helping the team develop pre-trained models and task-aware models that accelerate the analysis of pre-clinical drug development based on omics data to identify the genes, pathways, and targets that lead to disease phenotype, drug response, cellular differentiation, etc. My line team is also in charge of developing causal inference technology and data extraction technology used on medical claims data to provide corroborating evidence for drug effects in practice, including the open-source python package causallib.

Background: I am an interdisciplinary researcher combining machine learning and computational biology, with a proven track record of leading and assisting scientific projects. My academic experience includes physics, computer science, computational biology, and systems biology. I did my Ph.D. at the Hebrew University in Jerusalem, followed by a postdoctoral rotation at Mount Sinai School of Medicine and a junior faculty position at Columbia University. I accumulated extensive experience developing dynamic models, machine learning methods, and simulation tools and analyzing high-throughput sequencing data. I have more than 15 years of experience working closely with colleagues on the design of both hypothesis-generating and validation experiments while participating in both local and international teams.

More specifically, during my tenure at Columbia University, I developed algorithms to analyze gene regulatory networks that uncover how changes in these networks affect changes in phenotype. In this context, I created an algorithm called DeMAND to understand the mode of action through which perturbations (e.g., drug or disease) affect cells. Additionally, I created machine learning models to identify which features contribute to disease progression and mortality.

In two previous postdoctoral positions, one at the Hadassah Medical School in Jerusalem and the other at Mount Sinai School of Medicine, I analyzed the dynamics of small gene-regulatory modules, focusing on accurate and efficient simulations that take into consideration the stochastic effects arising from low copy numbers.