• Morphnet: fast & simple resource-constrained structure learning of deep networks (gordon2018morphnet)
• Continuous integration of machine learning models with ease.ml/ci. It would have been a better contribution if the code was open (I am not able to find it on their lab's github) and the project was ready to use, since similar ideas are tried by almost every ML practitioner.
• A bunch of Clojure stuff. Types, testing, concurrency. Also I feel like cookbooks (like Clojure's) make a lot of sense when the audience is not a beginner programmer.
• Model-based testing without models: the todomvc case study (bainczyk2017model)
• Inside Kdenlive: How to fuzz a complex GUI application?
• I am looking into testing methodologies for complex UIs and found a nice path where you learn an fsm-ish model of the system's behaviour and then, essentially, fuzz it. learnlib does that.

• Are we any closer to learning what sparks lightning?
• Finished The Psychopath Code.
• How Erasmus Darwin’s poetry prophesied evolutionary theory

Bibliography

• [gordon2018morphnet] Gordon, Eban, Nachum, Chen, Wu, Yang & Choi. 2018. "Morphnet: Fast & simple resource-constrained structure learning of deep networks", 1586-1595, in in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, edited by
• [bainczyk2017model] @incollectionbainczyk2017model, title=Model-based testing without models: the TodoMVC case study, author=Bainczyk, Alexander and Schieweck, Alexander and Steffen, Bernhard and Howar, Falk, booktitle=ModelEd, TestEd, TrustEd, pages=125-144, year=2017, publisher=Springer