Title:Machine Learning for Visual Scene Understanding
Posted: October 3, 2016
You’ll be responsible for designing and building both supervised and unsupervised learning technologies to parse objects and determine their state as far as it is relevant to driving decisions. States might include object classifications, traffic signal states, orientations, wheel angles and human actions.
You’ll have practical experience of using machine learning and it would be advantageous, but not essential, for this to be in the field of convolutional neural networks for computer vision. We’d also expect you to be familiar with some of the techniques for unsupervised learning. In any case, you’ll have a good working knowledge of at least one widely available ML toolkit, such as Theano, Caffe, Tensor Flow or Torch.
You’ll have proven software development experience using C/C++, Python, Scala or Java. You’ll be familiar with OpenCV.
You’ll have a 1st or upper 2nd class degree, or equivalent, in computer science, mathematics, engineering or physical sciences, with a strong mathematics content, from a top university.
Experience from the automotive or computer gaming industries and familiarity with open source development platforms, such as Linux and Git, would be assets.