DBTB INT Elad Ferber r

[Music] my name is a lot Ferber I'm the CTO of echo labs I think trying to codify our thoughts about how to use data in medicine that we gathered over the last almost three years that we're active in this air in this field and sharing them and seeing also how what other people are doing I think that's great I think you know we're living in a in an era where data is becoming so redundant and so available that you know it's actually pretty hard to try to separate what's good what's not i think it's there's a lot of a lot of opportunities out there to analyze data and to do it right and it really feels like gold rush even here in Silicon Valley everyone trying to trying to find new ways to analyze data in ways that create value to people enter companies so it's extremely exciting so many things to do [Music] that I think when you analyze data that will be used in decision making in health care there are a few challenges that don't exist when you analyze data in other fields and involving the clinicians inside the decision-making process that data science algorithms usually take might be useful also keeping things very simple and really delivering the results and the recommendations in the language in a format of you know medical diagnostic tools that already exist and I think that can make the barrier to entry a little lower because when you try to do diagnostics using data or trying to come up with a clinical recommendation that's something that it was usually usually done by humans and so there's a lot of you know the transition between a machine based algorithm and what were yours to do until now is we need to make it simpler and easier to swallow so I think that's the key lesson [Music] yeah I think that there are many people around here call themselves data scientists I think that's great because you know we use data everywhere so a data scientist could be an experimentalist it could be a programmer and could be a statistician but I think that you need actually to have very good basis in all three so you know without great programming skills and really understanding the latest tools and the latest methods it's hard to really get the best results you can and work very fast and large data sets you know statistics is extremely important and actually often overlooked a very solid statistical background and experimentalism so knowing how to get data how to source data how to run experiments what are the key endpoints that you need to test knowing how to design those I think good data scientists that we've been trying to hire and hired for echo labs you no answer all three and so I think it's really important to keep yourself well rounded when dealing with this problems because they're so open-ended sometimes [Music] you