Designed and implemented experiments based on the analysis of predictive models, leveraging machine- and deep-learning in Python alongside large-scale clinical EHR data, including unstructured text, to drive novel health research. Consulted on statistical matters across the company, particularly on clinical trial design and hypothesis testing, which included A/B testing. Built data processing and analysis functions into Python codebase.

Wrote and published technical papers to demonstrate novelty and significance of experimental results; developed technical aspects of grants to fund large scientific and engineering projects:

• Worked cross-functionally to construct research plans and timelines
• Managed and processed large-scale clinical EHR data for use in analysis using a Linux machine on the AWS cloud computing platform, MongoDB, and PostgreSQL
• Generated analysis reports and visualizations using Matplotlib/Seaborn
• Worked across teams, including engineering and sales, in order to communicate data needs and uses

Applied techniques include:

• transfer and semi-supervised learning,
• deep learning (RNN, CNN, feature extraction, GAN) (implemented in Keras, Tensorflow, PyTorch)
• NLP (Doc2Vec)
• supervised learning (tree ensemble, generalized linear models, regression analysis, SVM)
• unsupervised learning (clustering, dimensionality reduction)