Senior AI Engineer (VSLAM)

Your Mission

This position is responsible for designing, developing, fine-tuning, and optimizing deep learning models for object detection, recognition, segmentation, and pose estimation, as well as human activity recognition. Additionally, by incorporating reinforcement learning for robot teaching, he/she will develop innovative AI solutions for robotic applications.

What To Expect

• Have strong interests in the following domains, including but not limited to:
  • 2D/3D object recognition, segmentation, and pose estimation
  • Feature extraction (e.g., SuperPoint)
  • Reinforcement learning for robot teaching
  • Human activity recognition
• Design, develop, and optimize deep learning models (e.g., convolutional neural networks, transformers) for object detection, recognition, and segmentation using 2D image data and 3D point cloud data.
• Implement and optimize keypoint detection and description models (such as SuperPoint) and integrate them into various applications such as template matching, 3D reconstruction, and SLAM.
• Leverage reinforcement learning algorithms to develop robotic manipulation strategies through simulated robot teaching.
• Utilize deep learning techniques for human activity recognition in video data.
• Evaluate and calibrate various sensors to ensure the collection of accurate data for training.
• Pre-process and prepare image and sensor data for effective model training, including data cleaning and augmentation.
• Document technical reports, program codes, and setup manuals.

What You'll Bring

• A Master's or Ph.D. degree in Electrical/Mechanical/Computer Engineering or a relevant discipline with more than 5 years of industry experience.
• Strong foundation in deep learning techniques, particularly convolutional neural networks (CNNs) and transformers.
• Familiarity with object detection, recognition, and segmentation frameworks (e.g., YOLO, Faster R-CNN, Mask2Former).
• Experience with human detection and human activity recognition models (e.g., I3D, C3D).
• Familiarity with deep learning models for interest point detection and description (e.g., SuperPoint, LIFT, D2-Net).
• Understanding of reinforcement learning algorithms and experience with simulation platforms (e.g., MuJoCo, Isaac Sim).
• Expertise in deep learning frameworks (e.g., TensorFlow, PyTorch) and 2D/3D computer vision libraries (e.g., OpenCV, PCL).
• Proficiency in programming languages such as C/C++ and Python.
• Knowledge of data quality requirements for different AI training tasks and data acquisition protocols.
• Experience with data preprocessing and augmentation techniques for image and 3D point cloud data.
• Proven experience in developing and deploying deep learning models in real-world applications.
• Any patent, publication, specialist certification, or award in robotics and automation is an added advantage.
• An individual who is:
  • Professional and agile
  • Open-minded and seeks mutual growth with the company
  • Reliable and able to work in a team with high integrity
  • Goal-oriented and has a pragmatic “Can-do” attitude