Data Scientist | Deep Learning Researcher | Competitive Data Scientist
I am a Data Scientist with an MSc from Ben Gurion University. My expertise is built on a foundation of academic research and a track record of driving impact within early-stage startups.
Beyond the industry, I am an active competitor in global Data Science challenges, consistently achieving top rankings on platforms like Kaggle and MAFAT.
Joined as the 7th employee at Quago, a startup leveraging sensor-based data to detect fraud and anomalous behavior in mobile gaming.
Analyzed sensor data from over 100 million users, processing billions of events daily using PySpark.
Developed production-grade anomaly detection algorithms from scratch, focusing on minimizing false positives.
Led client-facing POCs and internal projects for detection and tooling, enabling deeper analytical insights.
Engineered an automated, multi-agent swarm system leveraging the Kimi LLM to autonomously tackle and solve complex Kaggle competitions. The platform features a fully integrated website UI and a dedicated Telegram bot for seamless, real-time interaction and monitoring.
View on GitHub
Spearheaded a comprehensive research initiative to detect sophisticated bot networks.
Developed a novel unsupervised graph grouping algorithm to identify complex user behavior patterns.
Led the project end-to-end, from initial problem discovery to production deployment - driving deeper analytical insights and successful client POCs.
Ben Gurion University
Dec 2023 - July 2025
Focused research in the field of Deep Learning and Computer Vision.
Completed my thesis under the supervision of Prof. Tammy Riklin Raviv.
The Open University of Israel
Feb 2021 - July 2023Focused extensively on Machine Learning and Deep Learning courses.
Dean's List (2021, 2022)
Achieved 2nd place out of 338 teams.
Implemented a sophisticated feature set and a large-scale ensemble to classify device behaviors from anonymized, label-agnostic web activity logs.
Awarded a prize of 10K dollars.
Finished at rank 19 out of 1,412 teams.
Developed a Deep Learning architecture using WaveNet and GRU blocks to classify multi-agent behaviors from noisy keypoint data.
Finished at rank 19 out of 616 teams.
Implemented a novel Graph Neural Network (GNN) architecture using dual embeddings and residual connections to optimize AI model compiler configurations.
Finished at rank 64 out of 1,876 teams.
Extracted behavioral features from keystroke logs and built a diverse ensemble of MLP, Gradient Boosted Trees, and Tabular Transformers to predict essay quality.
Interested in collaborating on deep learning research, competing in hackathons, or discussing data science opportunities? Reach out!