How are you recruiting? I find references and local recruiting to be the most effective.

not much, just waiting for my next ticket to be assigned. Reading tech blogs and papers in the meantime.

unfortunately this is how it works in real life too. Stay savvy!

we're all on google server's registers at any point ;)

@RememberMe

Paper and pencil is king when I need to actually think. Whenever I read a paper or try to figure out a concept, I use paper and pencil. However, after that step, I like to transfer those handwritten thoughts to my computer for convenience.

I literally "work" 4 hrs a day, usually in the morning. After lunch my productivity drops. I spend the rest of the day on "light" work like emails, code review, or reading research papers.

@billgates Yes, I agree with your point on how most of the time we only see algos used in technical interviews. Your day to day will probably revolve around for loops, arrays etc.

The first question is if you are looking for a job in Deep Learning. If you are, I would recommend learning maths / read papers over knowing the TF API because the field of ML is distinct from software engineering. It values theory / fundamentals over knowledge of libraries. From what I've seen, it's hard for software engineers to break into deep learning positions without some research / math background.

If you're not looking for a job and are just playing around, I would still say having knowledge of maths will make it much easier for you to pick up new concepts and papers. I personally learn "top down", where I first do a project, and learn the background maths at the same time. However, my footing in linear alg, probability, and calculus have made it much easier to understand concepts in the field.

I would say you have to know the math / background knowledge to do anything cutting edge with tensorflow / pytorch etc.

If you're just looking to do something simple like image classification then it is possible to get by with just blogs and tutorials. There are many tutorials where you can just run a script on some data and get the output. If you want to go beyond just being a "script kiddy" I would recommend fast.ai for a course with a good mix of math and applied coding.

It's like the difference between a programmer who knows algorithms / data structures versus a person who learned web dev directly. Both may be very adept at what they do but having a deep foundational knowledge will allow the programmer to recognize and solve more interesting problems.

Chrome takes up more ram than others browsers but it usually isn't a problem. If other applications need the RAM chrome will relinquish it. Chrome is faster than other browsers and it's heavy use of RAM is one factor.

@antoni4040
Matt Damon's attractiveness bypasses all genders, sexualities, ideas, and concepts.