Advance human motion understanding at Mecka AI through foundation models, human-scene interaction, and large-scale vision training. Work model details are not specified.

As a Research Scientist, you will investigate full-body kinematics, human locomotion, and interaction in complex environments. The position emphasizes proprietary model development, severe occlusion handling, and temporal learning from continuous ground-truth motion data. Your research will support perception needs for embodied AI and robotics.

Key Responsibilities:

- Architect models for full-body mesh recovery and pose estimation
- Develop biomechanically constrained losses and temporal regularization
- Track bodies through complex allocentric and egocentric spaces
- Map contact, torques,



and environmental affordances
- Prototype action segmentation, intent prediction, and sensor integrations

Requirements:

- Expertise in articulated tracking and human body pose estimation
- Solid theoretical and practical parametric body-model knowledge
- Experience with inverse kinematics and dense mesh estimation
- Proficiency with PyTorch and distributed training frameworks
- Experience with massive human motion datasets

Bring human motion research, physical modeling, and scalable deep learning skills to Mecka AI’s robotics-focused perception work.

📌 Human Motion Research Scientist (Richmond Hill)
🏢 mecka
📍 Richmond Hill