Design and implement advanced deep learning architectures to enhance perception capabilities within ADAS systems.
Maintain and continuously improve existing models by optimizing performance, addressing issues, and refining architecture and algorithms.
Perform detailed root cause analysis of production issues and develop sustainable, high-quality solutions.
Optimize model performance with a focus on latency, efficiency, and resource utilization for real-time embedded deployment.
Integrate and validate deep learning algorithms on automotive-grade hardware and embedded SoCs.
Collaborate closely with data engineering, data annotation, and platform engineering teams to ensure smooth data flow and seamless model integration.
Provide regular updates and technical reports on model development, maintenance progress, and performance metrics to management.
Qualification
Advanced degree (MDoctorate (PhKey Benefits & PerksComprehensive Health
Required
5+ years of professional experience developing, training, validating, and deploying deep learning-based perception models for ADAS or related computer vision applications.
In-depth understanding of training and inference pipelines, including data loading, augmentation, and loss function design.
Advanced degree (M.S. or Ph.D.) in Computer Vision, Robotics, Machine Learning, or a closely related discipline, or equivalent industry experience.
Strong proficiency in Python and a deep understanding of software design principles and development best practices.
Expertise in PyTorch (preferred) or TensorFlow for large-scale model development and experimentation.
Practical experience with data pipelines, distributed training, and machine learning experiment management tools.
Proven ability to work effectively in a collaborative, cross-functional team environment.
Preferred
Comprehensive understanding of machine learning algorithms, including classification, regression, and clustering methods.
Experience deploying and optimizing models for embedded or automotive SoCs (e.g., NVIDIA Drive, TI TDA4, Qualcomm Snapdragon).
Proficiency in model optimization techniques such as quantization, pruning, and knowledge distillation.
Doctorate (Ph.D.) in Computer Science, Artificial Intelligence, or related field is a plus.
Strong programming experience in Python and/or C++ within Linux development environments.
Familiarity with automotive perception workflows, datasets, and evaluation frameworks (e.g., KITTI, Waymo, Euro NCAP)
