Shape embodied AI perception at Mecka AI with 3D computer vision, PyTorch, and human-scene interaction modeling. Work model details are not specified.

This Research Scientist position owns the human-scene perception loop for Mecka AI’s data engine. You will build proprietary architectures, train on massive multimodal datasets, and solve emerging algorithmic challenges as robotics hardware and customer needs evolve. The scope includes motion, contact, physics, and deployed robotic control connections.

Key Responsibilities:

- Train proprietary foundation models for human body tracking
- Handle energetic motion, self-occlusion, and crowded scenes
- Adapt research to unblock robotics product capabilities
- Integrate physical contact, gravity, and momentum reasoning
- Deploy models for emerging perception challenges

Requirements:
- Deep learning and 3D computer vision expertise
- Large-scale vision model training experience
- Mastery of PyTorch and deep learning scaling
- Experience handling multi-terabyte image and video datasets
- Relevant publications or human motion dataset experience

Use scalable computer vision and physics-aware tracking to develop Mecka AI models for embodied robotics systems.

#J-18808-Ljbffr

📌 Computer Vision Scientist Embodied Ai (Richmond Hill)
🏢 mecka
📍 Richmond Hill