Back-end Clojure Developer.Implemented REST api routes for serving data from Amazon RDS and ElasticSearch sources.Wrote code for interacting with external APIs (e.g, Facebook, Twitter, YouTube, Instagram, Vine, Salesforce) to pull data into our analysis and storage pipeline.Implemented computational models in Clojure, Torch, and the probabilistic programming languages STAN, and Anglican.Collaborated with our front-end team to fine-tune our internal API for their needs. Worked in concert with our software architect to design and implement our back-end systems within the Amazon AWS infrastructure.Worked with our VP of Product to determine avenues of research which would generate useful business insights, including descriptive and predictive models.

*Generative 3-D models -- in-browser editing and stochastic sampling of curves and surfaces.. Implemented Gaussian Process sampling over vector functions of meshes in Javascript.. Implemented Rational Gaussian curves and surfaces in Javascript.. Used THREE.js library for in-browser display and editing of stochastic objects*Kaggle Data Science competitions.. Clojure: Yandex Web Search, Allstate Purchase Prediction.. Python: Traveling Santa Problem, Observing Dark Worlds*Wrote Restricted Boltzmann Machine library in Clojure*Wrote basic web server in Node.js for parsing and storing of 3-D object files. Ran on Heroku.com*Generative News.. Wrote web scraper for News headlines and implemented docNade network in Clojure

-- Computational and empirical studies of multi-sensory fusion --*Designed and conducted experiments for assessing sensory fusion in neurons and behavior*Modeled brain activation by sensory inputs using Generalized Linear Model regression*Assessed multi-sensory fusion in neural populations via Maximum Likelihood decoding*Evaluated models via hypothesis testing, including Monte Carlo methods*Modeled neural learning as density estimation using undirected graphical models (RBM)*Modified and implemented Kalman Filter Mixture Model for semi-supervised clustering of non-stationary neural data*Identified and removed data pipeline bottleneck: designed and implemented custom application (in Matlab) to automate data pre-processing. This resulted in 10x reduction in person-hours required.*Performed and taught neurosurgical procedure for intra-cortical electrode array implantation*Presented results in lectures, posters and journal articles*Mentored graduate students

-- Spatio-temporal analyses of the neural control of movement --*Recorded and analyzed activity of large neural populations during arm movement*Statistical analyses of the relationship between neural activity and movement*Performed, taught, and improved implantation of high channel-count electrode arrays*Presented results in lectures, posters and journal articles*Guest-lectured on methods for extracting and using brain signals

*Trained customers (scientists and surgeons) on neuroprosthesis implant procedure*Co-wrote manual on implant procedure*Contributed to data required for initial clinical trial of intra-cortical neuroprosthesis