Responsibilities

• Dataset and Evaluation: We focus on curating high-quality datasets tailored to autonomous driving scenarios and designing robust evaluation metrics to accurately assess algorithm performance.
• Active Learning: Our team explores techniques for efficiently selecting and labeling informative data points, minimizing labeling efforts while enhancing model performance.
• Network Architecture Search: We investigate methods for autonomously discovering optimal neural network architectures, specifically tailored for label generation from sensor and video data in autonomous driving contexts.
• Transfer/Low-shot/Long-tail Learning: Our efforts include developing strategies to leverage knowledge from related tasks or domains, addressing scenarios with limited labeled data (low-shot learning), and handling class distribution imbalances (long-tail learning) commonly found in autonomous driving datasets.
• Efficient Learning and Inference: We optimize learning algorithms and inference processes to ensure resource-efficient utilization, crucial for real-time deployment in autonomous driving systems.
• Privacy: Our team prioritizes the development of privacy-preserving techniques, ensuring the handling of sensitive data while maintaining high-performance label generation, in compliance with privacy regulations and safeguarding user information.

Qualifications

• 5+ years of practical experience (Candidates nearing completion of a Ph.D. degree may apply)
• Master's or Ph.D. degree in Computer Science, Electrical Engineering, Mathematics, Statistics, or a related field relevant to machine learning, or equivalent experience
• Strong background in Linear Algebra, Probability, Signal Processing, and machine learning concepts
• Proficient programming skills in languages such as C/C++, Python, and others

Preferred Qualifications

• Experience in development related to autonomous driving and robotics, including Object Detection, Semantic Segmentation, Depth Estimation, and Transformer-based models.
• Experience in building and utilizing automated learning pipeline systems.
• Track record of publications or contributions in relevant fields, such as CVPR, ICCV, ECCV, NeurIPS, AAAI, etc.
• Enjoyment in discovering and solving new problems in the field.