The first is an end-to-end machine learning case study: building a model, containerizing it, and then deploying it with Azure Model Management + calling it as a REST API. The second is an exploration of the mathematics required to implement a neural network. The final example is two quick demos for Build 2018 - one on Azure Cognitive Search, one on Azure DataBricks.
Examples were generated in RMarkdown. The textbook are figures and code to support an open-source DSP textbook from Richard Baraniuk; the other are examples for an R package.
I was also selected to serve as a mentor for the Chevron Data Science Development Program - an elite group of prospective data scientists, mostly coming from the internal research group. Almost all of my pupils had PhDs - which was a tad bit surreal, but very fun!
Four this year: two on DevOps for Data Science, one on Machine Learning at Scale, and an Introduction to Tensorflow with Keras.
Mostly documentation, especially Azure documentation - but you’ll also see ethical modeling guidelines, a few NASA Space Math examples, and machine learning code snippets for VS Code. Have also added minor changes to docs for Magenta, TensorBoard, TensorFlow, and Keras.
Created during the rOpenSci Unconference; was invited to attend by the good folks at numFOCUS. It was a heck of a lot of fun!
A mobile application that directs Microsoft employees to the closest vacant parking spot in Building 43, based on vehicle type. Uses YOLO to detect cars parked in spaces.
I, uh, also resuscitate Apple IIs and had a small LLC for wearable electronics for a bit (electricute.me). 😊