AI Engineer (Vision-Language-Action Models)

About the role

The AI Robot Association (AIRoA) is recruiting an AI Engineer to work on Vision-Language-Action (VLA) models and multimodal models for robotics. AIRoA is collecting one million hours of humanoid robot operation data with hundreds of robots and building an open robot data ecosystem where datasets, trained models and benchmarks are available to researchers worldwide.

Key responsibilities

• Develop Vision-Language models or equivalent multimodal models aimed at robotics applications
• Fine-tune existing models, run evaluations, perform error analysis and improve performance
• Build training pipelines with real-world data, design evaluation metrics and run iterative improvement cycles
• Prepare and preprocess image, video, language and action data, and set up training environments
• Research the latest research and technology trends and select technologies to improve models
• Set up training, inference and experimental environments for real-world deployment
• Work with software and robotics engineers to define requirements, design validation plans and drive development

Required profile

• Experience leading machine learning models from deployment to improvement and operation in a production service environment
• Experience implementing, training and evaluating ML models with Python and PyTorch
• Hands-on experience fine-tuning Vision-Language models or equivalent multimodal models
• Experience building training pipelines with real data, designing evaluations, doing error analysis and running improvement loops
• Able to turn the latest research and technology trends into model improvements and practical products

Required skills

• Python and PyTorch
• Preferred: VLA or multimodal model development for robotics
• Preferred: robot control, ROS / ROS 2, C++ and real-world hardware evaluation
• Preferred: sensor integration, actuator control, action generation and low-level control
• Preferred: simulators, Sim2Real and domain adaptation
• Preferred: cloud infrastructure (AWS or GCP), Docker, CI/CD and MLOps

Work location: Tokyo Ryutsu Center A Bldg., Heiwajima, Ota-ku, Tokyo.