Machine learning is on the verge of transforming healthcare, and the MGH & BWH Center for Clinical Data Science (CCDS) is at the forefront of this revolution. We are a fast-paced start-up embedded in two of the nation’s leading research hospitals, backed by industry partners like Nvidia and GE Healthcare. We have access to millions of medical records, an on-prem GPU cluster, and a top-tier team from industry and academia. We work closely with clinicians to solve critical problems in patient care – our goal is to make real products that make a real difference in the hospital.
The focus of our ML science team is to build models to solve critical needs in healthcare.
- Classification to expedite treatment decisions, prioritize worklists, and improve accuracy
- Segmentation and localization to improve the accuracy and speed of abnormality detection
- Volumetric assessment of pathologies, reducing clinician time per study thereby freeing them to focus on higher-value tasks
- Time series analysis to predict outcomes based on streaming patient data
Required skills
- Fluent in Python
- Highly comfortable in one of Tensorflow, PyTorch, or Caffe2
- Highly comfortable in the theory and practice of neural networks
- Capable of debugging faulty network architectures
- Experience distributing training across many GPUs
- Highly comfortable in traditional machine learning algorithms (e.g., SVM, boosting, bagging, etc.)
- Highly comfortable in SQL
- Comfortable working independently and defining measurable, achievable goals
- Experience shipping ML solutions in product
Nice to have
- Publications at top-tier ML conferences (NIPS, ICML, ICLR, etc.)
- Experience writing production-grade software (code review, unit testing, integration testing, CI, etc.)
- Experience programming in C/C++/Java/Go
by via developer jobs - Stack Overflow
No comments:
Post a Comment