How Data Scientists Can Build an Online Brand | CareerCon 2019| Kaggle

welcome back to career Khan everyone I am Jessica and I'm here with someone you've probably all seen right here on YouTube before suraj rebel popular YouTube machine-learning educator and director of the school of AI as well as Rachel Chapman our very own prominent data scientist and developer advocate at cackle they're both you two talk about how people such as yourselves can build an online brand even if self promotion isn't something you're used to or comfortable with they will also share how having an online presence can help you reach your career goals so Raj is going to share his experience first take it away Suraj thanks Jessica it's good to be here thanks everybody for showing up it's an honor to be here with kaggle and with you guys so I thought more than anything what I could do to help you out is to just give you an overview of my youtube journey so far for context three years ago I was working at a company in San Francisco called Twilio which offers a communications API to developers and my role there was to be a developer educator so what I was doing was I was writing developer documentation in about eight or nine different programming languages like PHP and Python and it was a lot of work for sure and I learned a lot in that time I was there for about eight months and one thing that I learned in particular was that there wasn't enough developer documentation in video form and there were a lot of at the time three years ago YouTube wasn't I mean it was big obviously but it what it continually grows every year and at the time there just wasn't enough so I so I saw an opportunity to start creating educational content around specifically around my first video was on Bitcoin but what I my passion was in artificial intelligence a field that I kind of I mean I took some classes at Columbia which I dropped out by the way I'm gonna college drop out I took some classes but I never feel like I never really processed it fully like I saw the equations I saw the slides and I kind of got I got the high level concepts but I didn't I didn't really get the math to be honest so what I started doing was I started making videos on artificial intelligence and machine learning and what I found was that by making these tutorials I was learning much faster than I would just like staring at a you know blog post or staring at a textbook and so for me creating educational content was a way to learn and I found that to be very valuable but more than anything else I had a mission and so a lot of people on social media which I guess you could call influencers vloggers streamers a lot of them are showing lifestyle they're showing how they live they're doing things that are entertaining generally and entertainment for the sake of entertainment is fine that's that's a fine mission but but what I felt was much something much deeper than that I didn't just wanted to be on YouTube to teach people about a ayah I had a goal from the very start and that goal was to make sure that as a society we can create artificial intelligence that would be maximally beneficial to every human on earth because I saw a trajectory of this technology getting better and better over time and I also saw the centralization of this technology into a few key powers and I felt like the way to prevent any kind of you know apocalyptic scenario where you know AI is manipulating people and taking advantage of them and exploiting them the way to get around that would be to make sure that as many people as possible understood how this technology worked and so I was just watching Eric Schmidt on Lex Friedman's podcast and and one thing that he said that really stuck out to me was that everybody has a six month plan or a one year plan but nobody has a five year plan and and and that's what that's what made Eric Schmidt to quote Larry Page the adult in the room when he got hired to be Google's CEO and that's probably a huge reason as to why Google was so successful because they were looking so far ahead and for me that was inspiring and just the idea of planning ahead so from the start I wasn't just try was making videos that were like they were funny that were entertaining I was kind of kind of like this gesture kind of figure but it was never about me it was about the technology and I considered myself to just be a vessel a martyr you know I can be that person because it's not about me it's about the technology and about the mission so my five-year plan was to win the hearts and minds of ten million developers so that I could transition the space from at the time what it was was mostly web app development mobile app development it wasn't focused on AI and I wanted to change the directive of the developer community and have it focused more on instead of sequential programming on learning and so that's a paradigm shift of thinking about software development in terms of learning and instead of giving the Machine steps to do something the instructions you give it the end goal the result and you will learn the steps to get there and that was that's basically but it's been a three-year journey to do that and I'm it's I'm not even ten percent of the way where I want to be but it's it's it's definitely been a lot of work so that's a little bit of context as to why I started making educational content and I've been doing it full-time now for three years YouTube has so now let me talk about platforms so YouTube has been my primary platform to do this I make these educational videos and each video takes around twenty-five to maybe thirty five hours of work to do and when I started I was doing all of the editing myself I was doing everything myself and it was a lot of work I learned a lot about editing and I didn't know how to use a lot of these tools I also just didn't understand audience sentiment as well as I do now because I have a very data-driven approach to content creation I'm not just creating content I'm looking at feedback I'm looking at audience feedback I'm reading every single comment I'm looking at the likes and dislikes ratio I'm tailoring my content to be maximally beneficial to the most amount of people and I don't I'm preaching to the crowd here but being data-driven not just in programming but in content creation is very import import it so that's something so YouTube has been my primary platform but I've also been using Twitter and LinkedIn and slack and Facebook and github as well many people don't understand that github is actually a social network there are 30 million now developers that live on github and you can follow developers and so now I have a good amount of followers which is awesome because the code that I write I meticulously making sure that the code is that it works now I'm not perfect obviously I have 429 repositories on github and I'm I'm always balancing quality versus quantity it's not like a solved problem for me but the goal is to make sure that a developer can just download that code and run it and it just works and if it works then they can start to build off of it they can get some ideas they can get better at it but that's kind of been some of the practice that I've learned from Twilio about code design and how to make sure that beginners it works for beginners so github has been a platform but YouTube has been my main platform when it comes to other platforms I use them primarily to promote my existing YouTube content now there are Instagram as well which is a newer one there are some cases where I am repurposing my YouTube content for other platforms so Instagram would be one example it's been quite a challenge to learn Instagram it's not exactly the most educational medium on the planet but I'm still learning there but what I've found is that using it as a feeder into YouTube has been I've been successful with that and taking clips of that content that are you know bite-sized and using that as a teaser that's been very good for Instagram so what else going to talk about here linked in so something I wanted so I feel like a lot of people don't recognize the power of LinkedIn as a platform for promotion self-promotion so the thing about LinkedIn is as opposed to other platforms it's very important to use keywords it's very important to use hashtags in particular because LinkedIn really promotes that and there was this there's this they might have fixed it by now but there's this hack in the LinkedIn algorithm where if you make a post and it requires the user to click on read more rap rather than being a shorter post LinkedIn will promote that more so there's all these little social media tricks that you can learn over time either by doing it or by looking it up just looking it up there are many medium blog posts on this but I think more than anything else consistency has been the key just being consistent with creating content with just being consistent I mean it's definitely been very hard to do I mean everybody's job here is very hard no matter what you're doing whether you're doing full-time software development back-end testing all of this stuff is hard but putting yourself out there for sure is hard putting your face out there putting your name out there and not just putting your face and name out there but attaching your work to your name and face is very hard because there's a lot of things that can go wrong there's imposter syndrome which I've definitely felt a lot of at several times I mean if you think about it this this this field is it's filled with PhDs which I love I respect you know I admire that it just wasn't the path for me so for me to come into the space without that credential in particular that's just a recipe for impostor syndrome right so but I was trying to prove a point that you don't necessarily need a PhD to implement this technology and I think that point has been proven but now the next point is to prove that you don't necessarily need a PhD to do research so with school of AI we're trying to grow the amateur research community in terms of social media promotion on that end now I'm promoting that much not just my own brand but the school of AI brand and so to segment my attention that way has been quite a challenge but what I found is that it's better for now for me to promote my own brand because that's where all the attention is as of now and what I'm basically a conduit for school of AI every word I type every tweet every video not just reflects on me but reflects on the community so it's quite a responsibility and I don't take it lightly so to summarize consistency is key to content creation when you the second point is to be data-driven about that content creation so look at the feedback take it take it in the chin yes get hurt because of course it's gonna happen allow yourself to get hurt allow yourself to feel these emotions but let them sit with you and learn from them right everything is a learning experience so data-driven content creation consistency and the third point is to have a mission have a mission that's more than just about promoting yourself or promoting you know a specific tool it should be something deeper and more meaningful than that something that will last you through these hard times and the last point is to have fun because this is hard for sure and if you're not having fun then you're not gonna you're not gonna stick through because you gotta be having fun whatever it is it's definitely full time content creation especially developer education is a lot of work and you gotta have fun so those are my four four tips for this livestream thank you guys for listening and now I'm going to hand it off to Rachel hey I made slides because my mind is just empty these days no it's too full and like things are falling out so if I don't write down what I'm going to say I will forget hi I'm Rachel if you've been tuning in to the morning sessions you might know me you might have seen me on Kaggle and I want to tell you a little bit of background about me my background is pretty different from Suraj's so I do have a PhD I don't know how much I use it in my day-to-day work but I do have a very fancy piece of paper that's on my walls every small should I look at it was like I was five years and my PhD is in linguistics specifically NLP computational linguistics so more on the computational side and during grad school I became more and more and more and more and more computational and by the time I left I was you know ready to be a data scientist and before that I was an undergraduate student also in linguistics you might notice a theme and in undergrad I decided that what I really wanted to be when I grew up was a popular science communicator about linguistics so like it's really cure Carl Sagan and Neil deGrasse Tyson not Bill Nye because he's not like a scientist this was a big discipline that he talks about and I was like that's what I'm gonna do and I started a blog and I started on Twitter and I posted you know blog posts about you know sort of intro linguistics things and I sort of kept that off kind of as I went through grad school and I also started talking more and more about NLP and data science and the things that I was becoming more interested in and I slowly really slowly started to grow a following so I don't actually know how many Twitter users I had before March 2017 because I just I didn't write it down the Twitter data only goes back three months and this happens to be the first time where somebody archived my Twitter profile on the wayback machine and in March 2017 I had 1,200 Twitter followers and in March of 2019 on the same day at I had 1,100 followers and that was in big part due to the fact that I started working at Kaggle and I had you know time to create more content and reach out to people and continued to interact with the community on there and I've also made a lot of really tough really dumb mistakes and I wasn't so much talking about being super intentional with branding with vision knowing the story that you want to tell knowing what you want to do and then doing it I did not do that I had sort of a if you like scroll way way way way back in my Twitter feed at some point it's like I saw a rabbit today here's a picture go into class steel of the building the class is in or something it's just more more slice of life and then as it became more of a professional platform for me I was like oh I should think about what I'm posting and how that reflects on me personally and how that might help me get a job and I did actually find this job my kind of job through Twitter because I met somebody at a conference and then he was like oh you should follow what if my advisees who graduated and I did and that was May crystal who has since left I go for Stack Overflow I mean if you're watching and she posted about this job and she was like hey I'm open for people Seattle is one of the cities you could be in and I was like I'm in Seattle I'm a data scientist so it was this sort of very organic connection I made just through meeting people at conferences and following them on Twitter and building in network and also this is the picture I use of me on literally every platform and I think I think it looks like me I know the lighting is kind of bad here so what is personal branding I think Suraj gave a really good introduction the way that I think about it is what is your public image when people think about you what do they think about and if you even if you haven't considered your personal brand you already have one right and you can't necessarily control it you can't control what other people think about you but you do have a lot of influence over it so in my early days of Twitter and blogging my personal brand that I was trying to develop was like hey linguistics popular science yay and at this point it's more like data science but also I want to cover technical content in detail I really want to spend time working through math and then help other people do that same thing so you can sort of think about some people that you might know in in the online space and I spent some time and I just thought about like three people so share she's really active on Twitter she uses a lot of emoji and she talks about politics a lot and that's how I think of shares online presence and her personal brand Siraj machine learning expert but makes it fun right I'm suppose this video really like slick and well put together and I expect sort of more of a produced experience and drinking machine learning expert educator maybe a little bit dry I think we've all we've always peaked at the machine learning course and but you know very sound technical content and this I just included a picture of a brand so those are you depending on your sort of comfort with English a brand is like the original use of it was like a stick that was on fire that you could use like mark things and then they added branding iron so these are really popular in the American West to brand cattle or livestock's you know who it belongs to and it's sort of very recognizable it's like oh I associate this goat or whatever with this person so some entomology kind of bring the linguistics in and I'm actually gonna have you guys do a quick exercise I have a timer nope that one's for 10 minutes actually went one for 20 seconds I want you guys to spend 20 seconds and we start the timer and I'm gonna sit here and stare at you and write what you want to be known for what do you want your personal brand to be so that could be your really good writing code and it's very easy to use and that's what you want people think about and they think about you or maybe you're a researcher and you work on optimization and that's what you want people to think about or maybe you make a lot of memes and you want people to think about you as like oh the funny a machine-learning meanie person so 20 seconds take some time notepad on your phone wherever write down what you want to be known for starting now it is alright that was 20 seconds and hopefully you have a little list of things or maybe one big thing that you want to focus your brand around so once you have an idea of what you want your brand to be what do you do with it so for me I think about my personal brand as a filter so to help me choose what I want to share online on whatever platform I'm using i'm going to talk most about twitter because that's where i'm most active but i'm Siraj mentions linkedin instagram I know a lot people use WeChat slack's whatever it is that you're sharing content online so my brand I mentioned used to be like linguistics that make it fun and these days it's I would define it as folksy non trivial data science so it's not super slick a lot of my stuff isn't super produced these slides that are just like black text on a white background is sort of the Rachel visual brand nontrivial I like I talked about the math I really want to make sure we understand what's going on in the code and I will take a long time to do that if necessary and the data science is sort of the general topic so having this in mind it lets me prioritize my time my time so Suraj's videos are beautiful he mentioned it takes like 30 hours to edit the video I don't edit my livestream videos because that's still it's still on brand for me if it's a live stream and it's just recorded and people can watch it back I would call maybe homes fun or informal as I would talk about my sort of visual brand identity if you will and also it helps me decide what to share so I have a lot of linguistic friends I still keep up with a linguistics literature a lot of my followers do not if Chomsky is trying to add a new you know operation to minimalism that's probably like absolutely nonsense to most of you so even if that's happening and it's sort of interesting and and my community is talking about it I'm probably not going to share it because it's not on-brand for my online presence that make sense I a lot of online communities and I've also done a lot of research on sort of social communities in my my dark days as a researcher not dark days but I don't really do that anymore and I sort of think of a taxonomy of ways of being in an online community whatever that community is so consumers are lurkers are people who go to a community receive content maybe you upload some things do you like something they don't produce a lot of content maybe you don't tweet if you're on Twitter and this is the type of account where if you have an account like this on a particular platform you're probably not going to be building your brand there you're just sort of gonna be receiving things so this is kind of how I use like Facebook like I look at pictures of people's dogs I don't paste to help post a whole lot and then there's aggregators so these are people who find useful resources and share it if you are trying to think about being an aggregator I encourage you to really think about what's the most useful and most novel thing you could share there's like you know a crew jillion people who have shared various data science cheat sheets on Twitter and they're not all necessarily gonna be useful for everybody and if I know see someone who's posted like six more all in a row I might unfollow them because that's not super useful for me right now and I think no more than two to three retweets a day if this is the sort of thing that you're focusing on and this will help you grow your brand but not as much as being a producer so producers create new content so you might produce YouTube videos like Suraj or I'm not writing a whole lot of blog posts these days but every so often I will I've read log tutorials and are early writing research papers but whatever it is that you're producing new things that you're writing or creating or filming or maybe you're doing a blog no podcast that's a sinking so you have voice recordings and whatever that is when you're making new content I would call you a producer and those are the people that I'm most likely to follow and the people who are doing the most to build their brand online I know that not everybody has time to like recorded a podcast in the evenings because maybe you have four children and a full-time job and that's enough that's absolutely fine I think this is just a helpful way to think about being a member of a community and how you fit into each of the communities that you're a part of and finally I have a good tweet in event week just to talk about I made a fake tweet just to give you an example of people who are producing content or aggregating content how you can maybe do a better job so this tweet here looking for hashtag ideas for hashtag projects to work on in hashtag data science help please this is a fake tweet I made it up but if I saw someone and this was in tweet that they tweeted probably I wouldn't follow this account so one thing is that it doesn't really follow the norms of the specific platform that it's written on so in Twitter especially in technical and machine learning Twitter people don't generally use a lot of hashtags Siraj mentioned so on Instagram people do use a lot of hashtags so if I made the Instagram post with some hashtags people might look at it and be like oh she doesn't really know what she's doing on Instagram she's not Instagram sorry LinkedIn those are both LinkedIn she doesn't really know what she's doing she's not a good member of this community she hasn't taken the time to learn the norms and people might be less willing to interact with me also for me as someone just interacting with this it's not gonna add much value to my Twitter feed I mean I guess sometimes I have projects that I was like oh cool somebody did that but I don't really have like a lot of ideas just sitting around that I can you know whipped up and even if I did this wouldn't necessarily be the best way for me to share them like if I was just like hey random person I'm gonna reply to you and I definitely talk back and forth with people on Twitter I'm not saying never ask for help but this doesn't necessarily drive me to interact with this person and it's also really focusing on what this person wants and not supporting the community so I was sort of a different example this is a real tweet it's my data science Renee and Renee is also asking for something so she says reply with your favorite recently published books I'm curious to see see hunter Owens and this is a quick tweet and the tweet she's quoting says what book slash resources etc should be on a getting started with Python or R for data science list I realize all my resources are three plus years old now so she's asking for something she's asking for recommendations of recently published data science literature but she's doing in a way that helps the community as a whole I'm she's curious she's helping somebody out there's some replies that people are talking about things have been recently published so it's useful for me and this seems like an account I do want to follow and I do follow Renee she's great because she is providing use for me as a someone who's interacting with her and also not all of her tweets are like this so she recently hosted a conference and she posted a bunch of slides and talked about what was going on in the conference and I've learned a lot of cool things from there so I think that's my last slide oh no I have one more slide with review so just review personal brand how you are perceived online you have some control over it especially with what you choose to share the more you do to support the communities that you're in the more likely you are to build an audience in that community so Siraj produces a lot of helpful videos in a space where there weren't a lot to begin with and that does a lot to support the community like I watch these videos I'm sure many of you have as well and just a final note I know I mentioned some numbers in here like numbers of followers I wouldn't recommend stressing about this unless it's part of your dog so astrologers contact creator and that's his job so clearly he cares about number of subs but if you're just interacting with people to learn things to make connections and I wouldn't necessarily focus on it and also follow a distribution on most platforms is highly highly skewed it's like the oh it's the statistical paradox where all your friends have more friends than you do rich get richer I think it's rich get richer anyway more than 95% of Twitter accounts of less than 500 followers so don't be like unless I have a thousand followers I'm not a real Twitter it's fine the thing that's really important is building connections making friends learning new things and being part of a community and it's those organic connections that at least for me have been the most valuable so that's what I have to say about branding I'll stop showing my screen now and I think either Jessica remark was gonna happen here I should have checked my notes sorry you're invisible there we go no longer invisible hi everybody thanks to both of you for all of your great points I personally really enjoyed the perspectives especially since you come from such different backgrounds but it really both like figured out your unique prominent brand and it works you know I think a key theme I learned from both of you is that there is no right badge that there is just cool so figuring out what your brand is and how you want to behave and position yourself within all the social media platforms out there today has proven to be very effective and a healthy career move for both of you so with that let's take a look at some of the questions from the audience our first question is from Timothy Timothy asks Suraj as somebody he had a full-time job in industry since doing Duty little time how successful would you say your brand is now and how many say or the careers might have reached out to recognize the work that you've done a great question thanks Sam so it's it's it's it's endless like hundreds a day yeah so I just I'm constantly eight percent of my engagement it's just no no no no no no I just want to keep making content and then sometimes I'll say yes if it's a cattle or if it's Nvidia or if it's Intel but I'm usually saying no so yeah like I don't even know how to the delta between where I was three years ago now is so vast it's not even I can't even imagine it's very different so so my point of saying that is not to brag but to show that if you make something that's quality and that's in demand that the intersection of those two things they will come if you build it they will come like a place to start from do you have any recommendations for a good platform to start for example creating vlogs on media creating an YouTube channel or first stick with portfolio projects and uploading like it definitely make content in general like as both a learning tool for yourself and to promote yourself and and I say this to my wizards might the people who follow my youtube channel not creating a portfolio isn't just about creating github repositories you have to sell those repositories and you have to communicate so making blog posts on medium is a great way to start because the production requirements are lower than it would be a video or podcasts these days we are now entering the golden age of audio and podcasts are everywhere and I haven't jumped on this trend yes let's see how that goes right now I'm very much enjoying YouTube as a platform because a skill that I have in particular is presented visually and you know whatever your skill is maybe you're better suited to just write or maybe it's to perform find out what that skill is that innate skill is and then just pick one and just go and then from there you can branch out later but just pick a platform and just roll with it rapid experimentation what are my values and next to our MVP attendee who has been here asking questions the entire day thank you so much for being here Lucy tan how do you balance learning with creating content when a lot of what we learn is already out there from tons of other resources and feel free to you know this question is open to both the views of the freedom whoever wants to speak first can all right yeah so how do I handle learning I have to look I have to learn a lot to really fast I felt I have to learn a lot of very advanced concepts that I'm not familiar with very fast and you know right now I'm spoiler alert making a video on things that I'm not familiar with related to react Jas and to you know different web development components I haven't necessarily built myself and the way I'm learning all this really fast is to watch these videos at 3x speed and I keep saying this a lot of people don't believe it but it's like lifting weights you know to start off it's gonna be hard but you just continually increase that speed point to 5.25 and your brain will adapt to that level of information processing and so I'm watching a lot of tutorials I am reading a lot of blog posts I'm synthesizing a lot from across the web and I'm building as well and you know I like to call it frustrated programming of doesn't work but that's you know you feel frustrated and you keep trying it but that's the learning process it's like Olympic athletes are so used to the pain when it comes to training their muscles we as learned learning athletes you learn to adapt to the pain of learning of not knowing a concept and to not just embrace it but - but - almost enjoy it you know weird I guess you could say masochistic way because you know in the end that you're gonna get something out of it so deal with the pain you know and just just keep keep learning my general process is a little bit more scattershot so in the morning I spend a good between 30 minutes an hour catching up on Twitter and I am pretty selective about who I follow on Twitter these days and I try to make it a good mix of NLP researchers because that's my area of expertise so I do unlike know what's going on in LLP and Saussure data science practitioners and also just like random folks cuz I like to keep it a little vibrant so seeing what the community is talking about is very important for me and then that tells me what I should spend my sort of limited time that I have to read papers and look at source code on the other thing that I do that's it's for you guys but it's also for me is we have the Cal Reading Club which is where I sit and read a paper aloud every Wednesday except not this one sticks I was doing something else and it's more interesting that it sounds maybe depending on how interesting you think that that sounds and having the forcing function of there are other people who also want to know this so I do have to sit down and do the reading because because they're gonna be there is very motivating for me as for how I pick papers mostly a quick papers that I think are currently the most important in the field of the field that I know the most about is that I'll be used they tend to be NLP papers so we read the just wanna transformer paper we read the paper we read the GPT - paper and things that are the community has sort of sort of picked up so deciding what to learn I use my community my network is a filter function I'm not just on archive every morning making time to learn I tried to do it efficiently so I'm doing something else if you're working in a full-time job and you're working on projects you have a little bit of flexibility ask for a project where you're gonna have a little bit of learning as part of it um if that's something that you can do so maybe people are having you may be you're currently working as a software engineer and you're implementing something and maybe you can ask if you have the ability to add a little bit of machine learning to that or if you're you're sort of creating data say hey can I build a data pipeline for this to send this to our data science team so trying to be efficient and incorporate learning into your work as much as possible continuing learning as you make content johnny mcgregor wonders do you both attend data yeah I do I'm maybe too many so probably my favorite conferences to go to just in terms of learning and research are the ACL conferences ACLs Association of computational linguistics so it's the big research organization for NLP and there's Knakal which is the North American one there's ACL which is the global one ACL which is the European one there's one in I just like pack scale anyway there's one in the Pacific so those are the research conferences that I'm most interested in for pure machine learning research I CML is a big one cvpr I think is more of a computer vision one but those are for researchers who are developing new algorithms for learning more from a practitioner side I really like the Python conferences so for data science specifically pi data is very good the PI cons and the regional PI cons tend to be more general because Python is a more general language um so you're gonna get a lot of like app development and web in front and back-end that may not be as relevant to you there as well but there is data science at the PyCon I'll be at PyCon and then for our the our studio conference is very very good use our which travels around is also very good and all of you our conferences are very data science focused because that's what R is for do you have other your favorites rush I think I'm always changing on this like I'm always adapting I think right now today I don't have a desire to go to any conferences because so the most recent one I went to was in videos GTC in San Jose great conference because I like in video but overall like an amazing as well the problem is that I'm so used to watching things at 3x speed similar reason why I dropped out of Columbia cuz I always always fell asleep in class because I was always you know probably ADHD whatever you have whatever whatever it is but I just my attention span is so limited that I when I if I do go to conferences its party and that's really it or speak but even that like just to like meet people and like have fun you know but the whole way track the hallway track exactly that's very much a track it's a track yeah and in terms of if I were going to go to another conference like probably I'm school that I was gonna have something but that's a really okay goal days is also very good a good conference that was cool yeah I mean there's so many a lot of different sort of like conferences but also I would say like developer and science meetups these days and I feel like that's sort of what both of you are alluding to there's a lot of different reasons to go to these conferences for Rachel it might be Rachel it might be to get up to speed with like the state-of-the-art NLP techniques or for Siraj it might be more to meet like newer figures in the field and really you know as we say the whole way track and sometimes I find myself at conferences as well doing one or the other and it doesn't have to be mutually exclusive depending on the situation and so I think getting and getting those names just now from both of you of the kinds of places you go to and conferences you go to that's really helpful well and then on the on the topic of sort of communities a user wonders should I focus on international communities or local and I think the motivation for this question is this you know to communities with different languages can be make it very different and also depending on where you want to be or where you're going to that that could also be some of the motivation for its question context yeah I'd say it depends on what your goals are I am in the very privileged position of having my native language also be sort of the lingua franca of academia and research I'd just say English there is absolutely a lot of value to creating content or even with formation translating content two different languages I don't want to be like English only English is the best you might get a wider audience but especially if you're looking to build a local community let's say in I'm trying to think of a country where English isn't like a major matrix language um I keep coming up with examples like Nigeria and big the - where a lot of people speak English Japan there we go a big Keiko community in Japan there's a big machine learning community of Japan and there's a lot of Japanese language content and community building that goes on in Japanese there and if you're in Japan I get in on it make some friends that was not an answer that was just me talking about languages it was an answer I wasn't a good answer oftentimes our data science community is very much separated by different kinds of languages and I feel like you know if I were in Japan and somehow I knew Japanese it would be a lot more relevant for me you try I would want to really meet people locally as opposed to you know just try to cater to a larger audience maybe that would be a later incentive but maybe not first yeah so I would add to that for sure like International is the way to go there's a lot of undervalued attention out there that I can see because I have the data of people who watch my content and they exist in Brazil and they exist in India area and they exist in a lot of countries that they're not think you know not necessarily english-speaking but they are undervalued and they're all coming up very fast and the job market is global and talent is distributed and talent is global and I think just I think this is something about the industry in particular our industry of data science and machine learning is that it still hasn't become you know accessible to enough people I think and I'm not perfect here I'm trying to become more accessible as well but we have to be speaking to a global audience because we're all this globalizing in and and that's just the future like we're all one big internet village so you know in terms of language translation I use transcription services like Rev to make sure that it's applicable to a lot of different languages but I may give like all sorts of little like cultural shoutouts to people like whether it's in you know anime culture or you know Indian culture or you know speaks in Spanish sometimes but you know it's just so people know like I know what it is like in terms of culture like I feel you you know it's not just it's not just America it's the world so yeah and I love being in data science personally for that same reason as well everybody is speaking you know when it comes to em all we're speaking the same techniques even if they may be in different languages great well we have one last question from Kelly Wong Kelly asks you're both seasoned content curators now but what was it like when you first started were there big challenges or mindsets you had to overcome any advice for people that are it's okay to teach people things you just learned so when I started I was like alright I've talked about linguistics we talk about the things that I know from every angle and that's especially for technical content sometimes the people who have just learned how to do something are the best teachers and content creators around it because you've just done the work of figuring it out and being like okay this thing is like this other thing that I know because I trained ponies or whatever and building those connections those allegories and then being able to take the work that you've done to understand something and immediately create content from that absolutely in terms of content any big challenges that you personally had to overcome when you first started perhaps more mentally as well as some advice for those that may be a little daunted only like when you start off you're gonna suck for sure like I do and it's always a work in progress and now I'm still trying to improve my skills content creation and I remember when I post one of my first videos to the e-learning subreddit the top comment was like never post here again I it's a nice little thing but like never comes and when my closest friends at the time you know sat me down he's like this is a project like this is you have a good job here at olio don't do this and so a lot of times people can't just sometimes you just have to trust your instinct and if you have intuition and it's backed by data not everybody can have that insight so you just have to trust yourself sometimes and that's all comes back to having that phishing if you have that mission it doesn't matter how hard it is or how many haters there are you just you just keep going really inspiring I just I just started my Twitter so now I'm inspired to become the next big I'd say in general when you are putting stuff out on the internet my personal best litmus test is what I want to give this content is there something I would be interested to see so just a little thing to take in your pocket absolutely and I think future literacy is a really important skill for us as data scientist to have to really practice practice looking ahead not just a year or five years but 20 years because reality is changing so fast and we're moving into an age where the absurd becomes normal and technology is the medium of making this reality happen and it's accelerating the advent of these algorithms are all compounding their impact on global societies so it's important to be future literate by understanding how not just current technology works but extrapolating to where it could be and if you do that you'll be a step ahead of the curve because the literacy is an important job skill not just in data science but in general so it's a part of being successful being able to predict where things are going and I think one of my friends just like on Twitter about how most CEOs are not future illiterate and that's the reason that these companies are not doing as well as they could be but a lot of the youth just by definition being more internet native or but this can apply to anybody at any age and it's just a mindset so try to be future literate pretty well I believe that is a wrap on both the fun Siraj and Rachel session thank you both surround Rachel for volunteering your time to be here today obviously and thanks everyone else for tuning in so Roger and Rachel will be heading to the career con flag channel now to answer more questions so feel free to direct your questions to them there and join us again very shortly at 2 p.m. EST for the final session of the day with one of the biggest names in AI and ml Jeff Dean here you

Channel: Kaggle

Views: 1,383

Rating: 4.5999999 out of 5

Keywords: Kaggle, Kaggel, coffee chat, live-coding, live, learn, api, cli, python, data, data science, interview, questions, transfer learning, coding, networks, programming, technology, tech, machine learning, AI, artificial intelligence, coders, programmers, help, tutorial, projects, 101, rstats, stats, statistics, what is kaggle, how to, github, developer, kernels, datasets, data visualization, deep learning, sql, challenge, competition, whitehat, code, lesson, CS, big data

Id: pZ5u_QSYfnc

Length: 48min 45sec (2925 seconds)

Published: Thu Aug 15 2019