It's easy to get excited about a new programming language but much harder to commit to using it instead of something more familiar on a real project. This talk will be a tale of just that -- I (James) will describe what it was like using Julia for all of the projects in UChicago's Machine Learning course. I'll go over what I loved and hated, what I learned about using Julia effectively, and how I think the language could be improved going forward while illustrating points with examples and livecoding.
I come from a Python background, so comparisons will be primarily to it, but the ideas are be relevant to anyone who uses dynamic languages for technical computing (R, MATLAB, etc.).