Responsibilities:
Supporting data selection related projects using data-centric AI technologies. This internship will mainly focus on OOD (out-of-distribution) Detection project for selecting unknown-unsafe corner case data to assist the data loop in the autonomous driving setting.
Applicants can focus on one of the following two tracks, or select both tracks according to the status of relevant projects:
Research track: engaging in AI research (e.g., on new OOD detection methods), including reading recent scientific papers in related fields, conducting experiments on GPU clusters, developing AI algorithms and writing academic papers.
Application track: engaging in applications and engineering work of SOTA AI methods to Bosch data, mainly for ADAS (Advanced Driving Assistance System) products. Conducting experiments on GPU clusters and online model deployment with model quantization.
Requirements:
Master or Ph.D. with computer vision experience, major in computer science, data science or related fields to artificial intelligence.
Proficient in Python and experienced in popular deep learning framework such as Pytorch.
Onsite working for at least 4 days a week.
Preferred:
Strong background in computer vision and related tasks (object detection, semantic/instance segmentation, lane detection, anomaly detection, OOD detection, etc.)
Have publications at NeurIPS, ICML, ICLR, CVPR, ICCV, ECCV, AAAI and other renowned venues.
Top ranks in computer vision related competitions.
Enroll as soon as possible, and could stay for at least 3-6 months.