The power of statistics and the recent breakthroughs in Machine Learning, especially in the area of deep-learning are really fascinating.
Here I’m playing around with the technologies available like the sklern
or tensorflow
libraries and see how they can be used to learn from different kind of data and what kind of applications are possible. You can find some IPython Notebooks (Jupyter) in this GitHub repository.
Right now, I’m interested in how you can make python work best for you in scale-up and scale-out scenarios and how you can get around the GIL issues if it comes to multi-threading and CPU intensive tasks. On github there is a small repository available Concurrent Python with some playground
scripts to explore the behavior and measure the performance.
While I was working on a GitHub Enterprise related project at work I noticed how much interaction is necessary to manage repositories via the Git Data REST API. That’s why I decided to write a library which is non-blocking (powered by aiohttp) and abstracts all this interactions / management tasks from the user. The aiogithub library is in very early alpha
stadium :) due to a lack of time.