Computer Vision Systems (Algorithm) Engineer

Job Description :

Successful candidates will be part of a systems and R&D team that develops embedded camera and imaging solutions, image signal processor (ISP), video codec hardware as well as advanced algorithms for computer vision and image/video processing.

Responsibilities :

The job responsibilities may include a subset of the following :

- Develop and/or optimize depth estimation and optical flow algorithms, enabling critical use-cases across mobile, AR/VR and automotive platforms

- Research and develop power-efficient computer vision algorithms, deep learning architectures and models in computer vision for AI-enabled camera, video, AR/VR products

- Developing or Optimizing image processing and computer vision algorithms for HW acceleration

- Support product teams for commercialization, such as solution optimization, performance profiling and benchmarking.

- Proficiency in programming languages such as Python (NumPy, OpenCV, TensorFlow, PyTorch) for developing computer vision algorithms and applications. Knowledge in C++ is a plus.

- Strong understanding of computer vision fundamentals including image processing techniques (filtering, morphological operations), feature extraction (SIFT, HOG, Optical Flow), object detection (YOLO, SSD, R-CNN), segmentation (U-Net, Mask R-CNN), and tracking.

- Experience with deep learning frameworks such as TensorFlow, PyTorch, Keras, and OpenCV for developing and deploying deep neural networks for computer vision tasks.

- Knowledge of popular deep learning architectures for computer vision such as CNNs, RNNs, GANs, & Transformer models.

- Strong mathematical foundations including linear algebra, calculus, probability theory, and optimization techniques (e.g., gradient descent, Adam optimizer) used in ML algorithms.

- Prior exposure to cloud platforms (preferably AWS) to deploy data pipelines and to use tools/services like AWS SageMaker, Kinesis, MLFlow etc.

- Exposure to Generative AI algorithms for Computer Vision like Stable Diffusion is a plus.

- Experience with GPU programming and optimization using CUDA or OpenCL for accelerating deep learning computations is a huge plus.

- Proficient in using computer vision libraries and tools for feature extraction, image registration and object recognition.

Required qualifications :

- Background in digital image processing and computer vision fundamentals

- Background in depth estimation and optical flow algorithms

- Working knowledge with Open CV and vision algorithms (feature detection/description and matching)

- Strong knowledge in data structures and working experience with C/C++ programming