DBTB INT Hiral Patel

my name is Hiro Patel I work for Yahoo and i'm a senior principal architect there i would say that there's a lot of talks on on data and then being able to share everyone's all seeing all the different experiences everyone has and how the same tools that you're using other people are using them differently and what are the pros and cons of all the things that they've had versus what you've had just seeing and hearing about all the differences about the same thing that you're using right so like what you're also trying to solve the same problems and how can you really solve them differently and what are the pros and cons of those different approaches that's a hard question i guess i've been working with data since I got out of college so I would say the coolest thing with working with data is there's a lot of open source tools that that make it easy to do you can even experiment on it on how to play with data on your own you don't have to be working for a you know fortune 500 company to get exposure to some of the tools like you had to when I first started nowadays you know you've got them you can do everything and on a Mac pretty much go download it try it out build your machine learning pipeline without having to you know pay you know thousands of dollars for software right so open sores in itself is the thing that's driving a lot of the initiatives in the big data space i believe my focus on in the talk was around doing prototyping I don't find your bottlenecks and and just addressing those you know there's also a lot of startups who tend to not look at the premature optimization concept as closely a little bit they just now let's move from one till to the other instead of trying to spend the time to actually understand the problem how are other people solving it maybe go to attend a few conferences right and and then go from there instead of oh yeah that's the you know one of the things i hear i've heard is oh yeah this we're using this tool we had a big prawns with the GC so we switched over to this other tool instead and it didn't have those GC problems and it worked great how much time did it take you to switch between tools instead of ten spinning maybe a day looking at how to optimize your GC parameters right so those are the things i think a lot of people don't get out and especially in startups you know don't make huge shifts until you really understand the problem and have done a little bit of prototyping right you know data science is something that like you know it's a new term I would say like when I was going to college you know I was taking classes on neural Nets but back then they didn't call it data science right so a lot of people with backgrounds and in computer science or like physics or math you know it's easy for them to just hit the ground running on the data science side I think the biggest problem that people have is making sure that they know how to code right so i think the coming of data science now with the right background even if you don't have a programming background is a lot easier to do because there's also a lot of open source tools built around it you don't need to know exactly how the Tool Works exactly except for you need to know how the what the inputs are and what the outputs are right so even if it's a black box it's fine so I think becoming a data data scientist is easier now than it was before and you know to me to become a great data scientist something I can't answer maybe Alexa will have a better answer because i am not a I'm not a great day of scientist I can't I can't give you any tips on that one you