Research Associate - 3D Vision for Robotic Pick & Place

ntu; Singapore; On-site; SGD 60,000 - 90,000; Full time

Job summary

NTU invites applications for a Research Associate to lead computer vision work in a Hyundai-NTU-A*STAR Corp Lab project focused on robotic pick-and-place. You will develop 3D object detection and segmentation models, create data augmentation strategies, and build algorithms to annotate 2D images using 3D data.

This role emphasizes high-quality experiments and rigorous reporting. Candidates should hold a Master's in a relevant field, have 2+ years of research experience, and demonstrate knowledge

Qualifications

• Master's degree required in a relevant field.
• 2+ years of research experience preferred.
• Deep learning expertise in 2D/3D object detection, segmentation and 3D vision.
• Experience with SAM, foundation models, diffusion models, and transfer learning is desirable.

Responsibilities

• Develop 3D objects and segmentation models for robotic pick and place.
• Develop data augmentation and synthesis to address limited data.
• Annotate 2D images via 3D data and ensure data quality.
• Review weaknesses of 3D detection and segmentation methods; experiment and analyze results.
• Collaborate with researchers and engineers; prepare reports and papers.

Skills

Deep learning

2D/3D detection

3D computer vision

Foundation models

Diffusion models

Transfer learning

Few-shot learning

Publications

Education

Master's degree in Computer Engineering/CS/Electronics or equivalent

Job description

NTU invites applications for a Research Associate to lead computer vision work in a Hyundai-NTU-A*STAR Corp Lab project focused on robotic pick-and-place. You will develop 3D object detection and segmentation models, create data augmentation strategies, and build algorithms to annotate 2D images using 3D data.

This role emphasizes high-quality experiments and rigorous reporting. Candidates should hold a Master's in a relevant field, have 2+ years of research experience, and demonstrate knowledge