AI and Synthetic Data Engineer-perception/physical-AI

Role purpose Develop synthetic-data pipelines and AI experiments for perception and physical-AI use cases using simulated industrial environments and sensor outputs.

Key responsibilities

• Generate labeled image, video, point-cloud, depth, segmentation, pose, and scene data from

simulation.

• Design domain randomization, scenario coverage, class balancing, and dataset versioning.
• Train and evaluate perception models using synthetic, real, and blended datasets.
• Measure domain gap and improve transfer through simulation, augmentation, adaptation, and

targeted real data.

• Automate experiments, metadata capture, quality checks, and reproducible releases.

Required experience and skills

• Computer vision or ML engineering plus synthetic-data experience.
• Python and deep-learning frameworks.
• Understanding of labels, ground truth, dataset bias, evaluation, and experiment reproducibility.

Preferred experience

Omniverse Replicator, Isaac Sim, diffusion models, YOLO/RF-DETR, segmentation, point clouds,

domain adaptation, or MLOps.