rdchambers.net

Welcome to my home page, and thanks for visiting!

I’m an machine learning engineer and a generalist software developer / data scientist. I have domain experience in deep learning, data analysis, computer vision, machine learning, and GIS (primarily in Python), and also in hybrid mobile app development (Javascript/Angular), server administration, and AWS deployment. My specialty is in solving scientific, numerical, or data-driven problems, and then in being able to deploy those solutions efficiently to whatever cloud-, mobile-, or embedded platform they need to live on.

I’m currently working as a Machine Learning Engineer at Facebook, and before that I worked as the technical lead for the Pet Insight Project, an exciting project at Whistle,aiming to improve the lives and health of millions of pets. Previously, I worked at Creare Engineering R&D, and then did consulting work after a pretty interesting stretch of travel, meditation, and work activities around the world.

Resume/CV

For information on my education and work activities, please visit my Resume, CV, and Publications page, or my Linkedin Profile.

Open Source Projects

Unfortunately, the vast majority of my work is proprietary. But, I hope to do a lot more open-source work in the future, and a few existing projects are below (see also my github repos).

• FilterNet Github Repo – This repo contains the code used for our FilterNet paper (a deep learning model for time series classification).
• Motorshed ( Github Repo ) – Motorshed is an open-source project to explore and visualize traffic patterns using beautiful, stylized maps. A Motorshed Map shows the flow of traffic to (or from) a single point on the map (origin point), from (or to) every other point on the map.
• Hypothetical ( Github Repo ) – A web app that helps you to explore how your cash flow situation might change in different hypothetical life events — e.g., moving to a new state, changing jobs, or having kids. It’s a work in progress, but it’s fun to play with. Written with Angular2 / Bootstrap 4.

• nbserve – A python library that lets you serve Jupyter/iPython notebooks as read-only HTML files over the web. One of the key differences between nbserve and other options (like nbviewer) is that it lets you run the files as they’re being served, allowing you to create dynamic, always-up-to-date ‘reports’. I wrote this to help bridge the gap between exploratory data analysis and creating up-to-date, client-ready HTML reports.

• OpenBCI Python GUI ( Github Repo ) My (now deprecated) Python GUI for the OpenBCI “Open Source Brain-Computer Interface”. I developed this in 2013 at Creare during a collaboration with the OpenBCI guys, and it was eventually open sourced. I think it’s a nice example of a concise Enthougth Traits-UI based GUI, and it contains an example of how to package a Traits GUI as a Windows Executable (see setup.py), which can be really frustrating.

• TabSINT ( Github Repo ) https://creare-com.github.io/tabsint/ I was the architect and lead developer for the original version of this software when I worked at Creare. It is the user interface side of a complete audiometric screening platform. It was further developed and then turned into this awesome open-source project by a great group of developers and engineers at Creare. Unfortunately, all of my commits were squashed before it was open-sourced. Uses AngularJS, Bootstrap, and Cordova.

Other Work

I also keep a page of old, random school work here, mostly because people who are searching Google seem to find the stuff useful.