Research Scientist - Physical AI & Robot Foundation Models

The role

This person will develop new RFM model architectures and learning methodologies. They will work on improving RFM pre-training and post-training performance and researching data mixing strategies to improve generalization.

What you will do

• Develop new RFM model architectures and learning methodologies
• Improve RFM pre-training and post-training (SFT, RL) performance
• Research data mixing strategies through collection and composition of robot learning data (Real/Sim/Web)
• Analyze latest research trends such as VLA / World Model and integrate/improve internal RFM
• Apply research and development results to products and services

Required qualifications

• Master's degree in computer vision, machine learning, robotics or related fields and 3+ years of related experience, or a PhD holder (including expected graduates)
• Experience in RFM (VLA, World Model, etc.) or robot learning (Imitation Learning, Reinforcement Learning, etc.) research/development
• Capability to implement and train models using deep learning frameworks such as PyTorch

Nice to have

• Experience in large-scale dataset construction and data strategy establishment
• Experience using simulation environments (Isaac Sim, MuJoCo, etc.) or actual robot platforms
• Experience in multimodal foundation model pre-training/post-training
• Publication record in top-tier conferences (CVPR, NeurIPS, ICML, ICLR, CoRL, RSS, etc.)
• Experience using large-scale distributed training infrastructure

Before you apply

• LG AI Research is not a military service designated company, so recruitment and transfer of technical research personnel is not possible
• The selection process includes a coding test and LG Way Fit Test (personality test)
• The selection process includes technical interviews (in-person) and Culture Fit interviews (in-person)

About the company

Physical Intelligence Lab researches and develops innovative Robot Foundation Models (RFM) to create world-class Physical AI technology.