Robotics AI Research Engineer — Multimodal Manipulation (Relocation)

Flexion; San Francisco (CA); In loco; USD 150.000 - 230.000; Tempo pieno

Vantaggi offerti da questo lavoro

Competitive compensation

Relocation & permit sponsorship

Open PTO policy & paid company curtail

Health, dental & vision coverage

Descrizione del lavoro

Flexion is seeking a Research Scientist/Engineer to push the frontiers of dexterous robotic manipulation by leveraging video demonstrations, wearables, and multimodal models. You will own end-to-end experimentation and deploy novel systems that turn cutting-edge research into real-world robotic capabilities.

You will design capture devices, curate datasets, and build scalable pipelines that translate human action data into foundation model and reinforcement learning workflows, advancing embodied

Competenze

• PhD in Computer Science or related field, or equivalent practical experience.
• Experience with vision, vision-language, video, and other multimodal models, especially ones trained with human data.
• Experience with multimodal generative modeling, training, and inference.
• Experience with reinforcement learning and imitation learning.

Mansioni

• Develop scalable data integration pipelines for human video demonstrations and wearable capture systems into foundation model workflows.
• Build automated pipelines for 3D object and environment reconstruction from capture data for simulation readiness.
• Create and deploy simulation environments with reconstructed assets for reinforcement learning and dexterous manipulation policies.
• Design and evaluate multimodal policy architectures translating multi-sensor data into robust robot control.
• Bridge human embodiment to high-DOF hands and humanoid platforms through novel algorithms.

Conoscenze

Vision models

Multimodal modeling

Reinforcement learning

Imitation learning

Datasets design

Formazione

PhD in Computer Science

Strumenti

UMI

Sensorized gloves

Descrizione del lavoro

Flexion is seeking a Research Scientist/Engineer to push the frontiers of dexterous robotic manipulation by leveraging video demonstrations, wearables, and multimodal models. You will own end-to-end experimentation and deploy novel systems that turn cutting-edge research into real-world robotic capabilities.

You will design capture devices, curate datasets, and build scalable pipelines that translate human action data into foundation model and reinforcement learning workflows, advancing embodied