Environment and background
Our focus at the Lawrence J. Ellison Institute for Transformative Medicine of USC is improving cancer outcomes by discovering optimal approaches to personalized therapy. Our vision is that a patient’s own metrics will determine their most effective treatment. They include molecular tumor markers, the dynamics of intracellular processes, or the behavior of cells across diverse microenvironments. We investigate a broad range of biological systems, including cell culture, mouse models, and patient-derived tissues.
We are seeking a machine learning expert who has a strong interest in cancer research. The candidate will collaborate with biological scientists and clinicians to understand their experiments and data to devise predictive machine learning platforms. Expertise in multiple machine learning approaches, including SVMs, DCNNs, and elastic nets is required. The ability to communicate effectively with biologists and clinicians on their own terms is essential. The successful applicant will be capable of working with complex multi-modal data and distilling them into formats appropriate for machine learning. Data sets will include histopathology images, high-dimensional multi-omic data, time series from live cell microscopy, and complex clinical data from electronic medical records. The candidate must be willing to perform low-level data manipulations, including locating/downloading data from public resources, populating databases, and data reformatting.
- PhD in computer science, statistics, mathematics, or related discipline.
- Three or more years’ experience in hands-on machine learning using real-world data.
- Expertise in:
- SVMs, DCNNs, elastic nets
- Cross validation techniques
- Model complexity and bias-variance tradeoff
- Extensive hands-on expertise with machine learning packages in high level languages such as C, C++, Python, R, Matlab.
- Experience with at least one of: Theano, Tensorflow, PyTorch
- Ability to effectively present technical concepts and research results to a scientific audience
- Effective technical writing skills
Key success factors in the performance of this position include a high level of attentiveness to detail, the ability to collaborate closely with others from diverse disciplines, flexibility in technologies, and enjoyment of learning.