Back to jobs

Research Scientist

Job description


Research Scientist $150,000 to $250,000 + Equity + Benefits + PTO
Remote or on-site - United States (New York or San Francisco)

Are you a Research Scientist with a background in generative AI models, looking to join a fast-growing start-up and work with cutting-edge technology in the AI video space?


This is an incredible opportunity to join a start-up with 8-figure funding, where you'll collaborate with world-class researchers and engineers to develop ground-breaking, lifelike avatars through advanced AI-driven content generation.


My client is a seed-stage company developing an AI-powered platform that transforms text, images, and voice inputs into dynamic, customizable video content. Their innovative technology is reshaping how users create and interact with digital characters across industries.


As a Research Scientist, you will focus on training state-of-the-art generative models for video, image, text, and audio. You will work with massive datasets, implement novel conditioning techniques for improved model control, and oversee distributed training across large-scale GPU clusters.


This is a fantastic opportunity to become part of a well-funded, high-growth company, working at the forefront of AI innovation while benefiting from clear progression paths into senior and leadership roles.


The Role:


  • Join a dynamic start-up with 8-figure seed funding

  • Develop and train generative AI models for video, image, text, and audio

  • Implement advanced conditioning methods to enhance model controllability and flexibility

  • Manage distributed training on GPU clusters using tools like Ray and PyTorch Distributed

  • Collaborate with top researchers and engineers to push the boundaries of AI capabilities


The Person:


  • Experience with Python and deep learning frameworks like PyTorch or Jax

  • Proven track record in training generative models, particularly for video and image generation

  • Experience with distributed training across large GPU clusters

  • Strong foundation in conditioning techniques for model controllability

  • Published research in top-tier conferences or journals