R&D Engineer – Computer Vision
An R&D engineer experienced in Computer Vision technologies, is being sought by Spectral Edge.
We are looking for an individual with knowledge of computer vision in general. Specifically: (i) image classification techniques (ImageNet). (ii) Object detection techniques (DarkNet, Yolo, SSD). (iii) Semantic segmentation techniques (Segnet). (iv) Knowledge of Neural Network frameworks (Tensorflow, Caffe).
On top of these techniques, the successful applicant will have the ability to create a training set, train a Neural Network and evaluate the results.
The key duties are:
• Applying computer vision techniques as appropriate to problems set out by the CTO
• Train a Neural Network for object detection and segmentation
• Work with software team to facilitate object aware image processing
• Implement Neural Network inference for customer environment in collaboration with software engineers
The key skills are:
• Strong Python, C/C++ skills and a working knowledge of Matlab
• Ability to optimise code to obtain maximum performance from compute resource available
• Knowledge and Experience of camera ISP architectures and/or video pipelines would be helpful
• Good mathematical background (so as to be able to gain an understanding of Spectral Edge’s IP). Masters degree or above in Math or Computer science expected.
• Good communication and team-working skills
The following additional skills would be an advantage:
• Experience of GPUs and languages such as OpenGL/CL, CUDA
• Driver development for cameras and sensors on Linux platforms
• Knowledge of ISP block algorithms, e.g de-noising or image registration algorithms
Candidates should be have a good command of written English and be able to work full-time at our offices in Cambridge, UK.
As well as a competitive salary, the successful candidate will be awarded stock options in an early-stage start up with considerable growth potential. This is a demanding technical role with a career path to more senior role as the business develops.
Spectral Edge is a specialist in Image Fusion, using patented algorithms known as “Phusion” to combine multi-spectral images or video into single images/videos in an aesthetically pleasing way. Applications include:
• Computational Photography on smartphones
• Video enhancement for content delivery, including assistive technology