Software Engineer in deep learning for medical imaging AI
This is an exciting opportunity for an enthusiastic software engineer with knowledge in deep learning to join the Wellcome/EPSRC Centre for Medical Engineering (CME) within an internationally leading university research School. The post-holder will be supporting researchers to develop and implement novel algorithms and pipelines for medical image processing using deep-learning and related AI approaches. The successful applicant will also be helping to promote coding best practice amongst researchers within the School and will take an active role in promoting open-source software supported by the School.
The aim of this post is to further develop the School’s flagship open-source deep learning platform: NiftyNet, a python package built as a TensorFlow-based framework and dedicated to medical imaging. This package is used by many users both within the School and externally. Beyond the CME, NiftyNet is also a key component of the software strategy for the newly funded London Medical Imaging & Artificial Intelligence Centre for Value-Based Healthcare, where it will enable, in partnership with NVIDIA and other collaborators, the deployment of deep learning techniques within hospital environments. The successful applicant will lead on refactoring and optimisation activities and will develop novel functionalities within the framework. The post-holder will also be actively collaborating with researchers to raise the quality of their implementation up to the expected standard for dissemination and technology transfer. The software engineer will work closely with other developers within the School to develop a modular software architecture in which the deep learning platform will work in synergy with other applications.
The successful candidate will have a graduate degree in computer science or a closely related field. They will be able to demonstrate good algorithmic and software development skills. They will also be able to demonstrate strong experience of TensorFlow or related libraries.