Experienced Research Engineer, Deep Learning - Medical Imaging at Bay Labs
San Francisco, California, United States
🇺🇸 (Posted Aug 5 2018)
About the company
Bay Labs applies deep learning technology to cardiovascular imaging to help in diagnosis and management of heart disease. We are an early-stage healthcare startup backed by top venture capital firms and the National Science Foundation. We created Bay Labs in 2013 to push the limits of deep learning to make an impact on healthcare. Our mission is to democratize medical imaging for greater patient outcomes and wellness worldwide. By improving access, value, and quality to medical imaging, we aim to promote and advance healthcare in both the developed and developing world. To achieve our vision of improving lives through better diagnosis and treatment of heart disease, we have assembled a team of domain experts in machine learning, visual neuroscience, robotics, physics, and cardiology. Our investors include recognized leaders in venture capital such as Khosla Ventures and Data Collective (DCVC).
Bay Labs is at the forefront of bringing deep learning advances to cardiovascular imaging to improve heart health worldwide.
Our mission is to make medical imaging universally accessible to improve wellness and healthcare. To achieve this, we’ve assembled a team of domain experts in machine learning, visual neuroscience, physics, medical devices, and cardiology. As an early-stage employee, you will have the opportunity to make tremendous impact while tackling critical unsolved areas in healthcare.
We are looking for an experienced engineer who has demonstrated capabilities to scientifically develop, benchmark, and validate a wide variety of deep neural network architectures for the purpose of extracting clinically-relevant knowledge from medical images.
Develop state-of-the-art and novel deep neural network architectures.
Develop training and testing pipelines to assess the performance of these architectures on clinically-relevant image processing tasks.
Keep up with deep learning literature in order to implement the latest techniques into our networks and pipelines.
Read relevant medical literature to be able to develop sound validation procedures/metrics.
Develop machine-learning algorithms on a breadth of software frameworks (Keras, TensorFlow, PyTorch, sci-kit-learn) and deploy on a diversity of hardware platforms (Titan X/Xp/V GPUs, chips for embedded applications like Nvidia’s Jetson TX2 platform).
Help disseminate theoretical and practical ideas in Deep Learning and Machine Learning to the rest of the team.
Skills & requirements
Master or Ph.D. in Computer Science, Physics, Neuroscience, Statistics, Mathematics or related fields.
5+ years of combined academic and/or industry experience in training, testing, developing, and analyzing deep neural networks (recurrent, convolutional, spatiotemporal, attention-based).
Demonstrated experience in implementing and tuning custom neural network layers and losses based on literature to extend capabilities of existing deep learning frameworks.
Deep theoretical and practical knowledge of machine learning principles and deep neural network techniques (like types of activation functions, pooling methods, 3D separable convolutional layers, use of inference acceleration frameworks, network compression/pruning/quantization techniques, etc...).Knowledge of computer vision and image processing techniques and methods.
Mastery of Python and the Python scientific stack (numpy, scipy, matplotlib, pandas, sci-kit-learn, sci-kit-image, etc...).
Advanced expertise with at least one major deep learning framework (Tensorflow, Keras, PyTorch, Torch, Caffe2, CNTK). Bonus: accepted contributions to one or more of these frameworks.
Experience with writing production code and the code review process.
Strong teamwork ethic, passion for learning, and desire to seek new challenges.
Bonus: Knowledge of C++ in an industry context is a plus.
Bonus: Experience in a startup environment is a plus.
Bonus: Publications in top-tier relevant journals or conferences is a plus.