Research Scientist (post-training)

What you'll do

• Lead research and develop supervised fine-tuning and RLHF pipelines for video generation models.
• Build evaluation and human-feedback data pipelines, and run experiments to assess alignment techniques and scaling.
• Collaborate with production and product teams, review research literature, and mentor junior researchers.

What they ask for

• Ph.D. in Computer Science, Artificial Intelligence, Machine Learning, or a closely related field.
• Strong publication record in top-tier conferences focused on reinforcement learning, alignment, or generative models.
• Extensive experience implementing and optimizing large-scale training pipelines using PyTorch; deep understanding of reinforcement learning, particularly RLHF.
• Experience with distributed training systems and large-scale experiments, robust evaluation frameworks, and complex shared codebases; strong software engineering and communication skills.

Nice to have

• Experience with diffusion models or other generative architectures, or fine-tuning large language or generative models.
• Experience with human-feedback data pipelines, video quality assessment, alignment techniques such as constitutional AI or debate, or perceptual quality metrics and human evaluation design.

Tools

PyTorch, NeurIPS, ICML, ICLR

AI focus

Alignment and post-training for diffusion-based video generation models, including supervised fine-tuning, RLHF, safety, evaluation, and human-preference feedback.