Research Scientist, Computer Vision MGenAI
Track this application
Get Started FreeMatch score against your CV
Get Started FreeTailor your resume to this job
Get Started FreeAbout interviewing at Meta
Recruiter screen, a technical screen (60-minute asynchronous challenge or paired live coding), a short culture-fit questionnaire, then a virtual onsite of about four rounds: fast-paced coding with multiple problems per round, system design scoped to level, a lighter behavioral, and an AI-enabled coding round assessing how you work with a coding assistant.
Read the full Meta interview process →Description
About
MGenAI is focused on making significant progress in AI-powered content generation for ads applications. In the past two years, generative AI has observed rapid advances, particularly in large language models and generative image models. Meta Monetization GenAI's mission is to advance the state of the art in applied generative AI technology and build products that will help businesses become more productive and efficient. The team is particularly focused on image and video ad creative generation.
Responsibilities
- Conduct applied research to advance the science and technology of image/video/text generation for applications in ad creative generation and editing
- Work towards long-term research goals, while identifying intermediate milestones
- Contribute to productionization of GenAI technologies in ads domain applications
- Influence progress of relevant research communities by producing publications
Minimum Qualifications
- Currently has, or is in the process of obtaining, a PhD degree in computer vision or machine learning. At least one publication as first author or co-author on NeurIPS, CVPR, ICML, ICLR, ICCV, or ECCV
- Experience in generative AI research (including PhD research)
- Experience in contributing to a team of applied research scientists and engineers in solving real world ML problems, in a role with emphasis on AI research
- Experience in developing and debugging in Python Experience in developing real-world applications using multimodal large language models
- Experience in developing diffusion models for image/video/text/multimodal generation with publications on this topic on NeurIPS, CVPR, ICML, ICLR, ICCV, or ECCV