DBTB INT Marek Kolodziej r

[Music] my name is Mark call jay i work at nitro and I'm a principal research engineer I [Music] think the questions i really like when people have recon create specific questions that they would like answered because sometimes i have the answer and sometimes i don't and even when I don't it's still a great learning opportunity [Music] as a person doing machine learning and artificial intelligence I kind of take it as a given but i think the exciting thing these days is what products going to be built with the data otherwise if the computers are just processing data without generating insights is just you know CPU cycles going and power being used up but no real value delivered so so I think at this point we actually have good enough data sets to make artificial intelligence more of a reality I think the key thing is that doing a linear algebra naively is bound to be quite inefficient and people spend a lot of time buying expensive hardware and if they're not utilizing it efficiently then there's no point in spending that money on the hardware when you could just use a better algorithm and have things run efficiently on the hardware that you actually have I would say this is pretty much the main insight well also the fact that you know the insights gained from doing linear algebra efficiently could translate into new algorithms that people are developing for which existing performant libraries do not yet exist so they could just take these insights and translate them to their specific problem [Music] I would say first of all we live in an open-source era and there are a lot of great repositories with code examples online that are open source it's not you know closed domain anymore and they're there plenty of really good courses on Coursera Udacity udemy etc so between that and a lot of really great YouTube dogs including those from by the bay and many meetups it's much easier than ever I would stay it's still risky because nothing really completely replaces a formal education and studying with with great experts is is really hard to do substitute but for people who already have some mathematical background it's probably comparatively easier I would say actually the thing about the this era about becoming a better professional is that these insights actually that can begin from from from the open source community and Open Knowledge community in general these may not get the person started necessarily because it can be really overwhelming to start from scratch based on just you know YouTube videos or Coursera but they can definitely help professionals grow and this is the case even for people with PhDs you know you just need to keep on learning because you're going to fall behind and at least the ongoing learning is really accessible these days that said you know we're at galvanize today and galvanize is a prime example that you don't you may not necessarily have to dedicate a whole PhD to to learn machine learning or data science you could start with something like the 12-week program that galvanized has and potentially the one year program just to get started get your feet wet and then you know you can keep on growing on your own as you're more aware of what's out there I think the biggest issue with ignorance is not not really knowing every single detail but being aware and not aware of what's even out there I think the goal of getting a person jumpstart is just getting the awareness of what's out there and then you can keep exploring on your own [Music]