About Me
Results-oriented AI/Machine Learning leader with a proven track record of driving innovation through holistic problem-solving, spanning foundational research down to user-facing production systems. Expert in identifying, defining, and mitigating complex AI risks with a delicate blend of deep technical skill, scalable systems architecture, and strategic deployment paradigms.
Core Expertise
Professional Experience
Harman International
- Spearheading the design and development of an innovative, multimodal content representation and recommendation system integrated into a user-facing product.
- Leading end-to-end machine learning pipelines, training architecture, and robust modular core systems.
- Driving the implementation of state-of-the-art transformer-based models and conducting original foundational research to close modern edge case gaps.
Meta Platforms, Inc.
- Owned cross-functional technical tracks for Responsible AI, automated content moderation, and algorithmic safety.
- Built a state-of-the-art Generated Content Detection framework reaching a 91% minimum precision-recall benchmark across diverse commercial generation models.
- Led proactive safety mitigation for Multi-modal LLMs, designing end-to-end guardrails to significantly drop adversarial downstream output risks.
- Implemented key algorithmic parity mechanisms reducing baseline bias in generative text-to-sticker spaces.
Alibaba DAMO Academy (Machine Intelligence Israel Lab)
- Developed novel attention deep learning frameworks optimizing automatic web-scale user photo album event classification and content parsing.
- Architected a custom automatic event tracking engine, enhancing overall parsing accuracy metrics (F1 score) by over 50%.
Elbit Systems, ISTAR Division
- Engineered classical and deep-learning target tracking, image registration, and object detection systems operating across multi-wavelength visual data streams.
Selected Publications & Peer Review
Professional Activities & Chairs
Organized the Workshop and Challenge on DeepFake Analysis and Detection (DFAD) in conjunction with CVPR 2024 (Seattle) and ICCV 2023 (Paris), facilitating benchmarks and research tracks around responsible generative modeling.
Education
Technion - Israel Institute of Technology
Focus area: Computer Vision research thesis; Full Academic Scholarship.
Technion - Israel Institute of Technology
Graduated through the highly selective IDF "Psagot" excellence track track.