Ali Arbabian

I am a machine learning researcher working on representation learning and large-scale content similarity at Meta. My work focuses on image, video, and audio representations, similarity search, self-supervised learning, and evaluation for highly imbalanced ML systems.

Previously, I built computer vision systems for checkout-free retail. Before that, I did applied research for deep learning in medical imaging.

Interests

• Deep learning representations for image, video, and audio.
• Self-supervised learning and multimodal embedding models.
• Similarity search, retrieval evaluation, and vector indexing.