CV

Ph.D

University of Oxford

Computer Science

M.Sc

Technion - Israel Institute of Technology

Computer Science

B.Sc

Technion - Israel Institute of Technology

Electrical Engineering and Physics

Single-Node Attack for Fooling Graph Neural Networks

Neurocomputing journal by Elsevier

Showcased that GNNs are vulnerable to a realistic single-node adversarial attack, even when the attacker cannot be chosen.

A Simple and Universal Rotation Equivariant Point-cloud Network

ICML TAG in ML Workshop 2022

A simple architecture which is equivariant to rigid motions with the ability to approximate any equivariant function.

Strategic Classification with Graphs Neural Networks

ICLR 2023

Learning in a setting where users that are dependant can modify their features to obtain favorable predictions.

Cooperative Graph Neural Networks

ICML 2024

A more dynamic and flexible message-passing paradigm in which each node can choose a different communication strategy.

Graph Neural Network Outputs are Almost Surely Asymptotically Constant

A new angle on the expressive power of GNNs by studying how the predictions of a GNN probabilistic classifier evolve.

Learning on Large Graphs using Intersecting Communities

A new and fundamentally different pipeline for learning on very large non-sparse graphs using intersecting communities.

Applied Researcher

eBay

Engineered large language models (LLMs) to match user queries with the most relevant landing pages for search engine optimization. This involved developing a novel keyword extraction technique, which I integrated with GPT-3 and subsequently fine-tuned for optimal performance. These efforts resulted in a significant improvement in user engagement, achieving a 30% increase in landing page click-through rates (CTR) within the first three months of implementation.

Data Scientist

SKF Group, AI Center of Excellence

Designed, customized, and implemented end-to-end machine learning pipelines, taking projects from initial ideation to production. I led research initiatives focused on event prediction, anomaly detection, and time series forecasting to support machinery fault diagnosis. My work included creating an event-based evaluation metric that enhanced the company’s anomaly detection accuracy from 78% to 90%, significantly improving operational reliability.

Algorithm Developer and Physicist

Rafael Advanced Defense Systems

A specialized military service assignment awarded to only one recruit per year, I created a neural network to predict $n$-body problem solutions for classified company use. Additionally, I engineered generative models to predict molecular dynamics pathways using physical simulations and developed new models to analyze stable and transient protein folding conformations, advancing our understanding of complex biological processes.

Tutor

Taught Math, Physics, and Computer Science for A-Levels, SAT, and college exams as an online part-time tutor.

Head Teaching Assistant

Head TA for Geometric Deep Learning and Graph Representation Learning courses, responsible over the other TAs, creating coding assignments and guiding 30 students through coursework.

Head Teaching Assistant

Head TA for the Geometric Deep Learning course, supervising 40 students and developing tutorials and code notebooks for the course.

Teacher

Taught Mathematics, Physics, English, and Coding for matriculation, SAT, and college exams, leading sessions of 30 students.

G-Research Grant - £2000

Graduated with highest honors (Top 1%) from a Master of Science in Computer Science

Clarendon Scholarship of over £200,000 towards tuition fees and living costs ( < 1% acceptance)

M.Sc Study Scholarship of over £45,000

Graduated with great honors (Top 3%) from a Bachelors of Science in Electrical Engineering

Graduated with great honors (Top 3%) from a Bachelors of Science in Physics