AI Research Scientist– Foundation Models & Embodied AI

ROBOSPEC PTE. LTD.; Queenstown; On-site; NZD 180,000 - 240,000; Full time

Job summary

RoboSpec seeks an AI Research Scientist to develop, train, and optimize advanced AI models for robotics and embodied intelligence, with a focus on multimodal foundation models and Vision-Language-Action systems.

You will work with robotics engineers to deploy robust, efficient models on real robots, explore state-of-the-art embodied AI techniques, and translate research into deployable robotic intelligence.

Qualifications

• PhD in Computer Science, AI, Robotics, ML, CV, or related field.
• Strong background in deep learning and large-scale model training.
• Hands-on experience with PyTorch and modern deep learning architectures.
• Experience in multimodal learning, foundation models, robot learning, or related areas.
• Strong research and programming skills.

Responsibilities

• Develop, train, and optimize large-scale AI models for robotics and embodied AI.
• Work on VLA, VLM, action models, multimodal models, and robot learning.
• Design and improve model architectures, training methods, and data pipelines.
• Fine-tune foundation models for robotic manipulation and real-world tasks.
• Improve model efficiency, robustness, generalization, and inference performance.
• Work closely with robotics engineers to deploy models on real robotic systems.
• Explore and prototype state-of-the-art research in embodied AI.

Skills

Deep learning

PyTorch

Multimodal learning

Robot learning

Research skills

Education

PhD in related field

Tools

GPT/LLM tooling

Job description

RoboSpec is looking for an AI Research Scientist to develop, train, and optimize advanced AI models for robotics and embodied intelligence.

The role will focus on multimodal foundation models, Vision-Language-Action (VLA) models, action models, and other embodied AI technologies, with an emphasis on turning research into deployable robotic intelligence.

Key Responsibilities

• Develop, train, and optimize large-scale AI models for robotics and embodied AI.
• Work on VLA, VLM, action models, multimodal models, and robot learning.
• Design and improve model architectures, training methods, and data pipelines.
• Fine-tune foundation models for robotic manipulation and real-world tasks.
• Improve model efficiency, robustness, generalization, and inference performance.
• Work closely with robotics engineers to deploy models on real robotic systems.
• Explore and prototype state-of-the‑art research in embodied AI.

Requirements

• PhD in Computer Science, AI, Robotics, Machine Learning, Computer Vision, or a related field.
• Strong background in deep learning and large-scale model training.
• Hands‑on experience with PyTorch and modern deep learning architectures.
• Experience in multimodal learning, foundation models, robot learning, or related areas.
• Strong research and programming skills.
• Ability to independently develop and validate new model architectures and training approaches.

Preferred Experience

• Vision-Language-Action (VLA) models
• Multimodal foundation models
• Robot learning and manipulation
• Imitation learning or reinforcement learning
• Diffusion models or flow matching
• World models
• Distributed training and GPU optimization
• Publications at leading AI, vision, or robotics conferences

You will work on RoboSpec's core embodied AI models, with the goal of building intelligent models that are not only powerful, but also efficient, reliable, and deployable in the physical world.