⏳ @DynamicWebPaige: BUILDING...
🗣 Notable recent conference talks:
- Deep Fashion: A Totally Rad Introduction to Transfer Learning (Anaconda Con, April 2018)
- Kill (Deep) Math! (ML4ALL, May 28, 2018)
- BUILD 2018 Azure Keynote: Azure DataBricks and Azure Cognitive Search (May 2018)
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.
- Attendees for the first two: ~100;
- Attendees for the third: 6000 in-person, 500k live-streamed, >2M in all.
📊 Digital Signal Processing with R
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.
👩🏫 Coursework for the Coursera study group I led at Chevron.
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!
In-progress examples of R code to supplement open-source statistics textbooks.
💻 Links to resources used for BUILD talks
Four this year: two on DevOps for Data Science, one on Machine Learning at Scale, and an Introduction to Tensorflow with Keras.
💾 GitHub contributions
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.
⚖ Proxy Bias Vignette with R
Created during the rOpenSci Unconference; was invited to attend by the good folks at numFOCUS. It was a heck of a lot of fun!
🚛 DeepParking: Microsoft One Week Hackathon 2018
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.
🐤 Some examples of Twitter rants:
-
Communicating complex topics + building an open-source community
-
Why there is often a disconnect between SWEs and data scientists
I, uh, also resuscitate Apple IIs and had a small LLC for wearable electronics for a bit (electricute.me).