 last year. So my name is Miles. That's me on Twitter, but I don't say anything very interesting and that's me on GitHub and I don't really publish anything interesting there either. I'm from England originally and it was interesting that panel discussion before talking about the first experience coding and I just realized how goddamn old I am. My first experience was probably when I was about 11, 1980. In those days it was you buy a magazine that had like a bunch of basic code. You'd be typing for like eight hours. End of eight hours it would just like print a circle on the screen or something, but that's that was my beginning. But yeah then from England I moved to San Francisco, worked in lots of startups and then on to Japan. That's Japan. I was looking for interesting pictures of Japan and that's the best I could come up with. So that's a really cool place and you know this is me on the way to work on a typical typical day in Tokyo. It's not really that was just some jobs that I think. And yeah I work at Cookpad and I've been there for about six years now which is the longest time I worked anywhere I think. We have a very simple mission to improve people's lives through cooking and in Japan there's a very big service and it's been going for 18 years now which is pre-google, pre almost everything. It's been a very long running service. I think somebody mentioned earlier that 80% of women between 20 and 40 use Cookpad every month which just doesn't seem possible but it's fact. You know we have a great business there but it's very you know very Japanese, very Japan-centric. Not many people including me can read any of this apart from the Cookpad at the top. So it's amazing for Japan but it's not very good for anybody else. So then we kind of embarked on a project to take what we've learned at Cookpad over the last 18 years collectively and bring it international and we started by acquiring a few other recipe companies in Spain, Indonesia, Lebanon and trying to make figure out how we can make cooking more fun for everyone all over the world. So we've gradually over the last year been pulling these other services and their content and teams into one platform, into Cookpad international and now we have about 25 million unique users around the world in eight languages I think that might be wrong. So we have these local teams and companies we acquired and we have a global development effort and I want to talk a little bit about some of the challenges and some of the tools we've come up with to try and make that a bit easier. Just to give some idea this is our HQ in Tokyo and you know we have a lot of experience like we said about how people cook in Japan. This is our office in, I'll just note to some in all these pictures. Yeah this is our office in Indonesia and what we find why we want people on the ground in different countries is unique challenges in each place and to have engineers living there and living the life, eating the food with their families is so much more powerful than me trying to figure out what people in Jakarta want to do and the problems they have. For instance one thing is that cell phones, cell phones are really really painful to use. Yeah actually seems great but they're always complaining but till I went there it was just mind blowing how you know you're just staring at a screen and the thing spinning everything you look at. So high performance is a really important thing and living there without every day you really want to focus on that problem. Then we have a nice office in Spain. Spain has a really interesting challenge that it's a very widely spoken language across 21 countries I think but the regions are very different so what we kind of present to people in Spain is very different to what people in Argentina might be eating, North South America and even North America with Mexico so we have a lot of challenges around you know trying to personalize the content, the same content in Spanish but for different audiences and having people in Spain and we have a couple people in there or someone in Mexico really really helps that. Again in Lebanon that's a crazy picture, it's even in Lebanon but obviously Arabic is right to left and we can't read that and it brings all kinds of challenges so having engineers and people on the ground there is really really important. Then we have a few other teams, Camicora, Palo Alto, coming up next in the UK so we're really really focused on this getting engineers all over the world to help us with this problem. So we've kind of developed a few workloads and we're still working and really love to hear anybody else's experiences but this is a few things that we've been working on which so let's start with this one. The first one is that kind of validating so everybody has tons of ideas we all think we know you know I think kind of the best thing everybody thinks they know the best thing when the team is very distributed that becomes very hard to communicate and we really wanted to enforce some kind of process and flow to not only come up with ideas but follow them through communicate the idea but then also validate and prove that my idea was actually a good idea. So we start with this hypothesis hypothesis development which is something I think we sort of thought about blogging about not really good and we kind of jumped head first in and this template kind of shows the main point so we really want to have an owner of this kind of card or story. It seems obvious but when you've got a hundred different things happening once there has to be one champion who's going to see that through one point person that you can talk to. We also want to know why you know why do we want to do this there must be some story behind it what's the reason for this feature or idea or change in the system. And then lastly is the hypothesis and this is a really this is really really difficult but really really valuable uh let's see if I can work this out here we go this thing. So in every story we try and come up with this little hypothesis we believe that something will result in the outcome and we'll know we've succeeded when there's some measurable results and it's seems trivial but it's actually really really hard to you have an idea that I want to make the uh I want to make the create recipe button purple because I like purple that's probably the kind of thing that might come up and you have to say well I believe that making the create recipe button purple will result in what what do I think it'll result in. Bad example but it's very hard you know so you have to really think through and think through what the the consequence is and also the success measurement will be about these slides. Yeah that's what we just talked about and that's uh that's something we kind of really abide by and it's it's really really hard to keep everybody focused and keep everybody in the company working on the same same kind of idea lots of people not just engineers as CEOs as community people lots and lots of people so it's it's really hard work. Um this one this is maybe my favorite uh and uh so Jam kind of coupled together by uh this guy in the front here and we really I really love using this thing and basically we're always making poor requests and we're always using stories from the developers we use give to tracker and countless times I'll be thinking you know um I want to work on a story to make this button purple and I'm think I'm about to start I'm thinking well what's the branch what should I call the branch name let's call it uh purple button or so I'm just I'm wasting time thinking of things um I'm just just busy work so this this flow kind of automates that from command line to the little tracker to github um and basically once we have all of our stories in people to tracker we just type get start on the command line and we see all the stories uh probably type in the number we want and uh pt flow will automatically create the branch and you can start working uh and then when you're finished get finished we'll push the branch and uh uh character github and create your poor request for you again it seems trivial but when you're doing this 10 times a day for six years you save a lot of I don't like to have to think so this means I don't have to think it takes all the information from the story title that someone else has already worked on so it's really really time saving the third thing that we also we I think caught this from someone at github uh Richard should have written his name by sandhime um so I'm presenting a ruby kaigi and this seemed like a really interesting thing that I would do at github ah that guy um and it's basically in cookpad japan we have an amazing framework called janko which is really involved and does amazing things for releasing small features to different people but we really wanted to start from the other end to make something super simple and we often have a situation where uh for instance in uh what's up here now maybe indonesia have some hypothesis from the validation board and they will create that hypothesis and they want to just test in one region so we'll use a feature toggle which is simple as something like this in a helper we'll have a a method that just says where it's enabled or this one's indonesia or for languages indonesian english and vietnamese and then in our code we'll just wrap the code with that um toggle and it seems a little bit hokey but it actually works it's really simple everyone can understand one of the keys is not to have hundreds of these at once so really to try and keep on top and this is something we don't do well but keep on top of them and make sure that the hypothesis for this trending keywords was proven and if it's proven we can remove this code and release everywhere the our idea is that we you know create a hypothesis prove it somewhere validate the results and then profit i think is the worst uh where are we it's going rather fast we have to talk more slowly uh this is another another also that pt flow i should have said is uh open source most of these things are all of these things are open source um the cp8 is a new uh a new project that basically we're trying to show that we care about really care about code and um push people to share what they're working on and as a working progress as early as possible so we encourage this kind of wip means a working progress which means i'm working on something but uh i kind of want some feedback i don't want to get too far you know into the down a rabbit hole without anyone else is looking at this so then we have this cp8 put pad box um and what it does it's basically just looking at our repo and tagging things and watching what's happening so when somebody puts uh uh wip in the title of their core request it will automatically label it as such in um uh in github which really helps simple to see visually and then when you remove wip from the title here when you're ready for a review it will remove that tag um another simple example is giving uh yeah giving a thumbs up will automatically tag it as reviewed um and that just means that basically once it's reviewed it's ready to be deployed by the the uh submit to wherever they want doesn't seem like much but when you've got you know 20 or so pull requests running at the same time to have this kind of visual easy glanceable thing uh really really saves a lot of time we all like saving time one of the new feature which i haven't really been paying attention to uh but uh basically we're scanning for old i think this is something that they do in the uh at rails um repository as well scanning for things that haven't been acted upon for four or five weeks then just closing them trying to smiley face or a heart or something but basically we're saying if you if you haven't got this done five weeks then probably not that important uh then yeah last last one i want to talk about is uh called company strano piesta and this is a little bit like uh it's again it's about communicating to a bigger team people distributed around the world in different time zones and we found that when we're releasing um a feature when we release it lots of people want to know but lots of people aren't looking at hub um and we want to wait a way to communicate that so we thought oh i know we'll have a release channel in slack and we'll type in what we've done but being you know humans we often forget to type that in or we forget something um so what piesta attempts to do uh and does really well actually is is is remove that kind of pain point and when you uh deploy to production it scans all the poor requests um since the last time that uh master was deployed and based on that it will open up uh vim console so you've done your cap strano deploy then you get this thing popping up that basically passes all of the the titles from poor requests since the last deploy and any images and you can you know quickly edit if you want but then when you save that it automatically pushes that out to the slack channel uh so now we you know you tend not to miss anything and everybody knows what's happening and we don't have to go back and think oh what do we do in this release what do we do in that release so it's a really great time saver so this is an example of uh little top pages broken on iE let's get past that one um uh the other thing yes it also creates a release on github for every deploy so it basically gives us uh automated change log and we can go back and see what happened so really really useful and all these things are kind of small in and of themselves but when you have a large team and we have maybe i think now in the international team uh the cookpad japan is very big and we have 100 engineers and tons of stuff going on international we kind of run separately and we have about 20 engineers in uh maybe six or seven different countries so spanning all time zones so this kind of over communication but over communication in a way that's familiar is really really important yeah but it's got a lot of facts but yeah so yeah wrapping up basically it's these these five processes uh you know evolved over the last year um and it's really really helped us um it's uh really key to helping us be productive but we're definitely interested in hearing what anyone else was talking about or doing in that kind of situation and also we'd like to just want to thank rexie and the team it's really amazing you know organization very new and very chaotic and i think everybody's really stressed out but so far really great really impressed and uh and it reminds me of ruby keggy a few years ago when it was still very new and people running around crazy and worried but yeah really going great so we're really happy to be involved and happy to be able to uh to contribute yeah so once again that's me and uh i was a little bit quick i should also mention a wicked tree uh compared with hiring uh so many things we're trying to do in so many places and it's it's really hard to find great people who kind of uh connect with our mission and want to help us so we're trying to find rails people ios android is going crazy in indonesia and middle east um machine learning also we're doing starting to get involved with the world image recognition and how that can help us provide tools and make things easier for people cooking and sharing recipes so yeah tons of exciting stuff um yeah what's about it maybe anybody have any questions about uh godzilla or anything i've talked about or uh living in england feel free to ask or so any any anything that you're doing similarly to to work with we'd love to hear yeah okay well thanks very much sorry for the rambling