CV

Expert Data Scientist

Franklin by QIAGEN

Joined Genoox (acquired by QIAGEN during my tenure) to build algorithmic solutions for complex genomic and clinical data. My role expanded into leading AI initiatives across the organization — I’m the go-to technical resource for AI development, LLM-based applications, and agentic workflows.

• Lead LLM-powered applications and agentic workflows end-to-end: identifying opportunities, prototyping, evaluation, integration, and production deployment
• Design, implement, and productionize machine-learning systems for complex genomic and clinical datasets
• Build NLP and text-mining solutions that transform scientific literature and unstructured content into structured, actionable insights
• Advise teams across the organization on AI solution design, technical feasibility, technology selection, and development best practices
• Optimize data-intensive algorithms with parallel and distributed computing
• Lead cross-functional projects with high standards for accuracy, testing, reproducibility, and production reliability

Senior Software Engineer

Chaos Labs

Chaos Labs builds automated economic security for crypto protocols — world-class risk simulations trusted by leading DeFi protocols such as Aave, Uniswap, and Osmosis to protect billions of dollars in digital assets.

• Built scalable, low-latency backend systems for the real-time DeFi risk platform
• Developed decentralized systems that let protocols incorporate dynamic, risk-reducing data directly into their smart contracts in real time

Computational Neuroscience Researcher (Yuval Nir Lab)

Sagol School of Neuroscience, Tel Aviv University

Used data-science tools to study human memory. My thesis developed MEGA (Memory Episode Gaze Anticipation) — a no-report paradigm that quantifies episodic memory from anticipatory eye gaze using signal processing, statistics, and machine learning.

• Co-authored ‘Anticipatory eye gaze as a marker of memory’, published in Communications Psychology (Nature Portfolio, 2025)
• Machine-learning classification identifies memory for an event from single-trial gaze features
• Research covered by Channel 13 News, The Times of Israel, and Neuroscience News
• In parallel, led the data-science pipeline for a sound-event-detection product at Sipple — model selection, training, and end-to-end AWS workflow

Senior Software Engineer, Azure Defender for IoT

Microsoft

Built advanced security solutions protecting IoT and OT devices at global scale on Azure, detecting alerts and generating recommendations from raw IoT telemetry across millions of devices.

• Designed and developed the Vulnerability Assessment feature — an algorithm matching CVEs to IoT devices at scale
• Built detection algorithms for security alerts and recommendations on global-scale IoT data

Software & Algorithm Engineer

Allot

Core engineer on DDoS Secure, a real-time network behavior anomaly detection (NBAD) and mitigation system, joining via a startup tiger team acquired by Allot.

• Re-architected the product backend from Ruby to Java, including database (PostgreSQL) and parts of the GUI
• Built a REST API for the BIRD routing daemon to mitigate malicious traffic during DDoS attacks via BGP
• Extended pattern-extraction algorithms to handle evolving attacks; applied multiple linear regression to predict traffic baselines under attack
• Won first prize in the Allot hackathon

Founder

Didact (MindCET EdTech Innovation Center)

Founded an online marketplace for video lessons and took it through the MindCET EdTech accelerator.

MSc

Tel Aviv University

Computational Neuroscience (Sagol School of Neuroscience, Yuval Nir Lab)

• Machine Learning (100)
• Computational Models in Brain Research (100)
• Brain Networks and Graphs (100)
• Systems Neuroscience (98)
• Bioinformatics in Radiology (94)

BSc

The Open University of Israel

Computer Science and Cognitive Science

• Data Structures and Algorithms (90)
• Linear Algebra (98)
• Probability (88) & Statistics (96)
• Brain and Cognition (98)

Anticipatory eye gaze as a marker of memory

Communications Psychology (Nature Portfolio)

The MEGA paradigm shows that monitoring anticipatory gaze with eye tracking can quantify memory retrieval without verbal report; machine-learning classification identifies memory from single-trial gaze features. Covered by Channel 13 News, The Times of Israel, and Neuroscience News.