The Data Scientist - Innovation reflects the mission, vision, and values of NM, adheres to the organization’s Code of Ethics and Corporate Compliance Program, and complies with all relevant policies, procedures, guidelines and all other regulatory and accreditation standards.

Data Scientists are responsible for curation and manipulation of datasets as well as data analysis and algorithm design. If the hardware/software engineers are the canvas & paintbrush, the data engineers/scientists are the artists – there exists a point beyond which they have zero engineering knowledge/background and vice-versa for the hardware/software engineers. The data engineers/scientists use their broad knowledge of applied mathematics and statistics to create new algorithms to interpret, manipulate, and derive actionable insights of highly multimodal clinical and operational data.

Responsibilities:

• Deeply iterative, collaborative design and development of algorithms directly with end users in patient care and healthcare operations
• Measurement of mathematical/statistical and compute/throughput performance of algorithms
• Dissemination and training of algorithms
• Maintenance and support of existing algorithms including monitoring for degraded performance and adversarial attacks

Required:

• Three to five years of relevant experience
• Experience with multimodal unstructured data (2D images, video, 3D/4D tomographic imagery, waveforms)
• Experience with multimodal structured data (SQL, NoSQL, legacy formats)
• Experience with “traditional” machine learning and data manipulation methods
  • Orthonormal decompositions (i.e. Fourier, Wavelet, etc.) of data
  • Time-series manipulations
  • Sparse data manipulation
  • Bayesian analysis
• Experience with “modern” machine learning methods
  • Gradient boosted trees
  • Deep neural networks and their frameworks (Tensorflow, Pytorch, JAX)
  • Bayesian neural networks (TF Probability, Numpyro, Neural Tangents)
• Optimization methods (gradient descent, Monte Carlo, and others)