AI Research Scientist, Computer Vision - Video Intelligence

Meta•Menlo Park, CA

4d•$122,000 - $181,000

About The Position

Meta is seeking an AI Research Scientist to advance computer vision research applied to Facebook Video Intelligence. In this role, you will develop and apply state-of-the-art deep learning and computer vision techniques to understand, analyze, and generate insights from large-scale video content across Meta's platforms. You will work alongside a multidisciplinary team of researchers and engineers to push the boundaries of video understanding, including temporal modeling, action recognition, scene understanding, and multimodal learning, directly shaping how billions of people experience video on Facebook.

Requirements

Currently has, or is in the process of obtaining a Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience. Degree must be completed prior to joining Meta
2+ years of experience in computer vision research, including deep learning approaches applied to video understanding, image recognition, or related visual domains
Experience designing, training, and evaluating deep learning models using frameworks such as PyTorch or TensorFlow for video or image tasks
Experience working with large-scale datasets, including data preprocessing, augmentation, and evaluation pipelines for visual data
Experience implementing and validating research ideas through rigorous experimentation, including ablation studies and quantitative benchmarking
Experience communicating technical research findings in written form, such as technical reports, research papers, or design documents

Nice To Haves

Experience applying self-supervised or weakly supervised learning techniques to large-scale unlabeled video corpora
Experience collaborating with cross-functional teams to integrate research models into production systems at scale
Research experience in temporal video modeling, optical flow, video generation, or multimodal video-language learning
Proven track record of contributions to peer-reviewed venues such as CVPR, ICCV, ECCV, NeurIPS, ICML, or similar conferences

Responsibilities

Design and implement novel computer vision and deep learning models for large-scale video understanding tasks including action recognition, temporal segmentation, and scene classification
Develop and evaluate multimodal learning approaches that combine visual, audio, and textual signals from video content
Build and benchmark video representation learning methods using self-supervised and weakly supervised techniques on large-scale datasets
Run controlled experiments to test research hypotheses, analyze results, and make data-driven decisions to guide model development
Collaborate with engineering and product teams to translate research advances into production-ready video intelligence features on Facebook
Contribute to the design and curation of large-scale video datasets and annotation pipelines to support model training and evaluation
Write automated tests and implement monitoring to ensure reliability and performance of deployed video understanding models
Communicate research findings through internal technical documentation, presentations, and contributions to peer-reviewed publications
Participate in code reviews to maintain code quality and share knowledge of computer vision best practices with peers
Identify and address privacy, security, and integrity considerations throughout the video intelligence research and development lifecycle