khiner

Given the "unreasonable effectiveness" of recurrent neural networks, and the increasing ease of their application to a given learning problem, I'm often left with the following uneasy feeling when reading through academic papers covering more specialized approaches: "Hmm, this was written in 200X... am I wasting my time reading this because deep learning could blow this technique out of the water?"

I want to develop my "smell" for judging whether reading through a paper for a non-RNN model is a good use of my time. Does anyone have any heuristics, red flags, smells, entire domains, etc., that they use to put things immediately in the category of "nah, this is smoked by deep learning techniques"?

I'm a software engineer, and I'm considering a career shift toward the space industry. I would love to develop simulations, or do web app development for internal tools or even for public-facing web resources.

From my brief research, it seems like SpaceX has developed a bad reputation for their poor work culture (see Glassdoor for examples). Has anyone heard otherwise? Has anyone heard that the work is rewarding enough to compensate for this? Are there any alternative companies with resources even remotely close to SpaceX that have a better work culture reputation? And finally (greedily), does anyone have insight on what would be most effective to focus on to prepare oneself for this kind of work if the only background they have is a typical Bachelors of CS degree courseload and several years of professional web development?

Thanks in advance for your insight and advice!