Transcript: DBTB INT Evion Kim r

[Music] my name is Evie um Kim I'm leave machine learning engineer at Matt Oberg so the interaction was really good so I've been to some other talks as well but the people's willingness to like know about what I'm talking about that that enthusiasm was like much higher than the average I would have expected I got a lot of questions and a lot of follow-up questions after like my tough so I had hard time coming back up here because there many people like asking customs and stuff so I like that level of enthusiasm data is so cool and exciting because the fact that we are using the data and trying to coming up with the better future is like not polluting it has been like that for like centuries maybe like even from the origin of species that has been the case because we learned for something from the our past history and trying to apply to the features that that's the natural thing but it can be done in much better faster scalable way these days thanks to all the devices we have you can collect more data we can have better understanding and we can apply it for the creating the better future in much massive scale with better accuracy so things that we have been doing so far can be done in much better ways that to accept what's exciting about data [Music] so the part that I wanted the part that I want to emphasize from my talk is that so there are a lot of talks about the machine learning about deep learning about the data infrastructure and what's the like better usage of it and stuff but how to actually using this kind of machine learning solutions on the rearward problem that connection seems to be less slightly loose connect compared to the other ones so I wanted to give a talk about that in a way that I had this problem i turned it into the machine learning problem by dice slice and dice it and applied like disqus inside for it so that's my intention of the top and whomever interested in like how to solve the problem using machine learning can get benefit from it I think [Music] so the term data scientist itself is kind of vague these days just like machine learning the big data like it's pretty buzzwords so whenever some company says they are hiring data scientist actually if you take a look at the job details like soon they are trying to hire it's all different someone is trying to hire the data and only someone is trying to hire the machine learning personnel someone's trying to hire the data infrastructure engineer so if you want to be data scientist and if you want to be great one one piece of advice that I can give you is have some capability on engineering itself it's really necessary so coming up with inside coming up with motor that's all great thing but when we are talking about industry not about academia and data scientists are the ones in the industry so you need to be able to ship things and make it production eyes and for you to do the products and I do need engineering power either have it yourself that would be very helpful if not you can when you're being interviewed for the data scientist the company which data scientists take a look at what the culture is like how the team is more mated formulated in a way that if that team itself has some other engineers who can build the things you ship you want to that'd be great so bottom line is in any way you need to have some engineering capacity that you can put your hands on that's the answer you [Music]