Research Scientist (post-training)

ZipRecruiter; San Francisco (CA); On-site; USD 120,000 - 180,000; Full time

Job summary

An innovative research lab is on the lookout for a talented Research Scientist to pioneer advancements in AI alignment and post-training techniques for video models. This role places you at the cutting edge of AI, where you'll lead research initiatives aimed at enhancing the quality and safety of video outputs. Collaborating with cross-functional teams, you will design robust evaluation frameworks and mentor junior researchers, fostering a culture of responsible AI development. Join this forward-thinking team and contribute to shaping the future of AI technology in a dynamic and supportive environment.

Qualifications

• Strong publication record in top-tier conferences focusing on RL and alignment.
• Extensive experience with large-scale training pipelines using PyTorch.

Responsibilities

• Lead research initiatives in alignment and post-training methods for video models.
• Design and implement RLHF pipelines for video models.

Skills

Reinforcement Learning

Alignment Techniques

Generative Models

PyTorch

Software Engineering

Communication Skills

Education

Ph.D. in Computer Science

Ph.D. in Artificial Intelligence

Ph.D. in Machine Learning

Tools

Distributed Training Systems

Evaluation Frameworks

Job description

Job Description

We are Genmo, a research lab dedicated to building open, state-of-the-art models for video towards unlocking the right brain of AGI. Join us in shaping the future of AI and pushing the boundaries of what's possible in video.

Role overview:

We are seeking an exceptional Research Scientist to join our team, focusing on alignment and post-training techniques for large-scale video models. In this role, you will be at the forefront of ensuring our diffusion-based video models reliably produce high-quality, physically accurate, and safe outputs that match human preferences and values.

Key responsibilities:

1. Lead research initiatives in alignment and post-training methods for video models, focusing on improved quality, reliability, and adherence to human intent.
2. Design and implement supervised fine-tuning and reinforcement learning from human feedback (RLHF) pipelines for video models.
3. Develop robust evaluation frameworks to measure model alignment, safety, and output quality.
4. Create and optimize data collection pipelines for human feedback and preferences.
5. Design and conduct experiments to validate alignment techniques and their scaling properties.
6. Collaborate with cross-functional teams to integrate alignment improvements into our production pipeline.
7. Stay at the cutting edge of the field by regularly reviewing academic literature in both generative AI and alignment.
8. Mentor junior researchers and foster a culture of responsible AI development.
9. Work closely with product teams to ensure alignment methods enhance rather than inhibit model capabilities.

Qualifications:

1. Ph.D. in Computer Science, Artificial Intelligence, Machine Learning, or a closely related field.
2. Must have:

• Strong publication record in top-tier conferences (e.g., NeurIPS, ICML, ICLR) with a focus on reinforcement learning, alignment, or generative models.
• Extensive experience implementing and optimizing large-scale training pipelines using PyTorch.
• Deep understanding of reinforcement learning techniques, particularly RLHF.
• Experience with distributed training systems and large-scale experiments.
• Proven track record in designing and implementing robust evaluation frameworks.
• Excellent communication skills with the ability to explain complex technical concepts to diverse audiences.
• Strong software engineering skills and experience with complex shared codebases.

Ideal candidate will have:

• Experience with diffusion models or other generative architectures.
• Background in fine-tuning large models or generative models.
• Experience working with human feedback data collection and annotation pipelines.
• Strong aesthetic sense and understanding of video quality assessment.
• Familiarity with alignment techniques such as constitutional AI or debate.
• Track record of successful collaboration with product teams.
• Experience with perceptual quality metrics and human evaluation design.
• Contributions to open-source projects in AI alignment or generative AI.

Additional Information

The role is based in the Bay Area (San Francisco). Candidates are expected to be located near the Bay Area or open to relocation.

Genmo is an Equal Opportunity Employer. Candidates are evaluated without regard to race, gender, age, religion, disability, veteran status, or any other characteristic protected by law. Genmo, Inc. is an E-Verify company, and you may review the Notice of E-Verify Participation and the Right to Work posters in English and Spanish.