Senior Computer Vision Software Engineer

We are currently looking for a Senior Computer Vision Software Engineer for Self-Driving Cars to work within our team.

We are a team of engineers that harness the power of artificial intelligence, video-based perception and deep learning to create industry leading automated driving technologies. We develop state-of-the-art computer vision algorithms, art neural networks based on video sensors with the focus on traffic/driving/parking scenes understanding. Our activities cover a broad set of computer vision algorithmic methods and entire data-driven development loop ranging from data selection and filtering to creating models for various vision tasks and to optimizing & deploying them on embedded automotive hardware, for self-driving / automated driving cars. Our applications cover various mobile platforms (vehicles, motorcycles, etc.), enabling automatic lateral and longitudinal control of the vehicles as well as high-definition maps creation.

In your role you will:

• Develop computer vision based environment perception algorithms for assisted and automated driving systems;
• Design, develop and optimize algorithms and software for single and for multi-camera video systems using a variety of state-of-the art methods (deep learning, structure from motion, optical flow, 3D reconstruction, clustering, segmentation, objects tracking, visual odometry, numerical optimization techniques, projective geometry);
• Create, adapt and train deep learning models that detect / classify / segment objects of interest in the context of automated driving, such as, but not limited to: vehicles of all types, pedestrians, road infrastructure elements, traffic lights, signs, free-space etc;
• Be responsible for the data-engineering related activities such as data filtering and automatic search, doing statistics, defining labeling guidelines, preparing the data requirements for each model etc;
• Be responsible for optimizing and deploying trained deep learning models and 3D perception algorithms to dedicated hardware devices;
• Evaluate and validate the performance of the models based on ground truth and reference sensors data, and will experiment ways to improve scenarios with undesired behavior;
• Evaluate and tune the algorithmic performance on large data-sets for reaching series-ready systems performance;
• Occasionally develop offline algorithmic methods for reference sensors data processing (e.g. LiDAR) and visualization software (e.g. 3D visualization);

Technology Stack:

• Python, C/C++
• Tensorflow, Keras, Pytorch
• Parallel computing frameworks (NEON, DSPs, OpenCL, CUDA, …)
• Horovod, Microsoft Azure
• Linux
• GIT, Jira, Bitbucket, Docker
• Jenkins, CMake or equivalents

Requirements

• You have at least 5 years of experience in the computer vision field.
• You have a university degree in computer science, automation, electronics and telecommunications, informatics, mathematics or comparable;
• You have a thorough understanding of computer vision / deep learning / machine learning (i.e., building blocks, mathematics, and intuition behind), algorithms, data structures and software design.
• You have an extensive experience with Python, object oriented programming and with deep learning training frameworks (TensorFlow, PyTorch, etc.);
• You enjoy working on both research & development related activities;
• You are motivated and committed to get involved, learn and contribute to the development of cutting-edge technologies in collaboration with a global network of experts;
• You have good communication, strong analytical skills, you are team oriented, have initiative, perseverance, and attention to details;
• You are self-organized and able to estimate, prioritize and schedule own tasks
• You are result and quality oriented;