DBTB INT Vinay Prabhu r

my name is Renee I work as a machine learning engineer with unify IV well to begin with this was my first talk that I gave as an employee of the startup that I joined a couple of months ago so that was the the first part of excitement secondly I think the quality of questions were also pretty mature and good and some of the interactions that I had after the talk so you had a kind of nicer like more thinking audience and I think the audience is kind of a differentiating factor about this particular even oh that's a good question to me a data puts bread on the table too piggish read I think beyond that it's just a artifact of human existence like it's basically that aspect that captures the reality right and reality is always fascinating and data if captured well should basically mirror that fascinating aspect and why and what is the core source of the fascination that's I think more philosophy and not a very good philosopher well firstly you know we work in the domain of seamless authentication so it's we have actually seen time and again that if you are able to just use the kind of sensors on your phone and you know you smart signal processing and machine learning algorithms then you can construct a pipeline based on which seamless authentication can be seamlessly realized just by using your phone as a passive authenticating device I think we will be able to solve the whole get past this whole password entrance should I say like a wicket gate which is kind of preventing people from you know using the information products that is being enabled in the society in a more seamless way so I think the big takeaway is that your phone is as we have seen from closed waters is loaded with enough sensors and there r NF idiosyncratic traits about you that I captured by the phone that you can the phone is actually a very potent you know seamless authentication device and we're basically you know on the path to make sure that that happens in the near future I think from I've see what I've seen is that a certain point of time people have realized that you know they can they have become a good data scientist I it's like one of those things you know when you go through a phase transition and at and post phase transition you realize that you are in a different pace but you don't actually get to realize that you're going through the journey like in my case my doctoral thesis basically dealt with a lot of real-world data and at a certain point of time you know I had basically worked on like computational political science data sets basically data sets coming from roll call votes in United States in it then united states census data then I work with data about restaurants in the bay area and about medical healthcare party claims so at a certain point of time I realized that irrespective of what domain the data is emerging from there's a commonality of math that can be thrown onto it and the commonality of pipeline that can be the data can be pushed through to get glean all these insights so as long as it a curious person you know this feels very organic and certain point of time I realized that you know I'd become a data scientist so the only session I can give is like dedicate at least one day a week to towards you know going through the state of the art literature specifically the flagship journals like GM LR and jossa and keep yourself up to date to me the most exciting thing is that everything happens so fast in the world of data sciences so you know what is relevant today can be rendered you know obscure in very short span of time so it's a very fast cycle industry so you have to you're constantly under this pressure to kind of keep maintaining your skills and it happens organically because it's so exciting right so just but I'd show that you know your ability to handle mathematical rigor doesn't go away and you're not seeing these algorithms as black boxes so basically diving deep into a what each function makes and don't be just in a function shopper on syfy or some fancy library without knowing the nuts and bolts of it so I think dedicating at least one day a week towards just you know getting your pen and paper and doing the math I think it's helped me a lot and something that my research advisor doing my PhD kind of instilled a tried lot to instill in me and I think that's paying off so that I guess one advice that I can give