Helix AI Engineer, Generative AI

Sourced from genai.jobs

Figure AI is hiring a Helix AI Engineer on its Generative AI team to build and scale generative models that power humanoid robot autonomy. The role centers on training and deploying diffusion and generative models across vision, video, and multimodal domains for perception, world modeling, and synthetic data generation. You will work closely with data, training infrastructure, and agent teams to integrate these models into the full autonomy stack. This is an on-site role at Figure's San Jose, CA headquarters, five days per week.

• Design, train, and deploy large-scale generative models, focusing on diffusion-based approaches for vision, video, and multimodal data
• Develop models that improve robot perception, world modeling, and prediction from raw sensory inputs
• Build generative systems for synthetic data creation, augmentation, and dataset scaling for robot learning
• Implement state-of-the-art techniques in diffusion, generative modeling, and multimodal foundation models
• Optimize large-scale generative model training pipelines across distributed systems
• Integrate generative models into the full autonomy stack alongside data, infrastructure, and agent teams
• Evaluate model quality, robustness, and generalization across real-world scenarios
• Design scalable experimentation frameworks for generative model development

• Experience training and deploying generative models (diffusion, autoregressive, or related approaches) at scale
• Strong understanding of modern deep learning techniques for vision and/or multimodal systems
• Proficiency in Python and deep learning frameworks such as PyTorch
• Experience with large-scale datasets and distributed training systems
• Strong experimental rigor and ability to iterate quickly on model performance
• Solid software engineering skills and ability to build reliable, maintainable systems
• Ability to operate independently and own ambiguous, high-impact technical problems

• Experience with diffusion models for image or video generation
• Experience with multimodal foundation models (vision-language or vision-language-action)
• Background in synthetic data generation or simulation for robotics or embodied AI
• Experience optimizing large-scale training (multi-node, GPU clusters)
• Familiarity with 3D, video prediction, or world models
• Prior work in robotics, embodied AI, or real-world ML systems
• Publication record in machine learning, computer vision, or generative modeling