Computer Vision / Deep Learning Scientist
Passionate about precision medicine and advancing the healthcare industry?
Recent advancements in underlying technology have finally made it possible for AI to impact clinical care in a meaningful way. Tempus' proprietary platform connects an entire ecosystem of real-world evidence to deliver real-time, actionable insights to physicians, providing critical information about the right treatments for the right patients, at the right time.
We are looking for Computer Vision / Deep Learning Scientist who are passionate about the prospect of building the most advanced data platform in precision medicine.
What You'll Do
- Research and development of novel imaging data based machine learning algorithms for the product platform
- Apply statistical and machine learning methods to analyze large, complex data sets
- Communicate highly technical results and methods clearly
- Interact cross-functionally with a wide variety of people and teams
- PhD degree in a quantitative discipline (e.g. statistics, statistical genetics, imaging science, computational biology, computer science, applied mathematics, applied physics or similar) or equivalent practical experience
- Experience developing, training, and evaluating deep-learning models using public deep learning frameworks (e.g. PyTorch, TensorFlow, and Keras)
- Experience developing, training, and evaluating classical machine/deep learning models, such as, SVMs, Random Forests, Gradient Boosting, CNN, FCN, ResNet, GAN, etc.
- Familiar with CUDA and GPU computing
- Knowledge of different medical imaging modalities, such as DICOM formats and pathology images
- Self-driven and work well in an interdisciplinary team with minimal direction
- Thrive in a fast-paced environment and willing to shift priorities seamlessly
Nice to Haves
- Kaggle.com competitions and/or kernels track record
- Experience with AWS architecture
- Experience working with survival analysis, clinical and/or genomic data
- Experience working with Docker containers and cloud-based compute environments.
- Familiarity with neural network techniques (batch-norm, residual connections, inception modules, etc)