SF Scala: Vlad Giverts Interview

hello everybody I'm alexic rubber off hello everybody I'm alexic rubber off the organizer and founder Joseph Scala the organizer and founder Joseph Scala and here we are at Clara the first time and here we are at Clara the first time it's our last meet up of the year and we it's our last meet up of the year and we kind of start with the new company this kind of start with the new company this is we're very excited and here we have is we're very excited and here we have vodka verts with us who is the head of vodka verts with us who is the head of engineering and Clara and he is a loner engineering and Clara and he is a loner friend and partner of subscribe he friend and partner of subscribe he hosted us before Thank You Vlad and hosted us before Thank You Vlad and we're happy to be here yeah happy to we're happy to be here yeah happy to have you guys here thanks thanks was fun have you guys here thanks thanks was fun thanks to the wetness so woman I think thanks to the wetness so woman I think at least two different previous at least two different previous companies yeah we met first identified companies yeah we met first identified you invited us and we shared the common you invited us and we shared the common law for intellij idea which was the idea law for intellij idea which was the idea of choice was Carla and still remains of choice was Carla and still remains the case right and then you where at at the case right and then you where at at work day mm-hmm right doing Simon doing work day mm-hmm right doing Simon doing machine learning and building the group machine learning and building the group there and so then you are kind of there and so then you are kind of inviting us to this new place so tell a inviting us to this new place so tell a little bit what is the trajectory which little bit what is the trajectory which is this called connection you know what is this called connection you know what is Clara what's no its new yeah so I is Clara what's no its new yeah so I mean what is Clara is probably a bigger mean what is Clara is probably a bigger question so I'll go into that probably a question so I'll go into that probably a little more separately ok but it just in little more separately ok but it just in a nutshell why does Claire exist and a nutshell why does Claire exist and that is the vast majority Polinsky do that is the vast majority Polinsky do not fully understand how our financial not fully understand how our financial asst to works yes and they are not asst to works yes and they are not empowered to own their financial lives empowered to own their financial lives and make the right kind of financial and make the right kind of financial decisions for their own benefit mm-hmm decisions for their own benefit mm-hmm so what we're trying to do is to empower so what we're trying to do is to empower them with kind of both education them with kind of both education information con transparency in our part information con transparency in our part mm-hmm to take control their financial mm-hmm to take control their financial lives and we're starting with mortgages lives and we're starting with mortgages mm-hmm scala connections interesting mm-hmm scala connections interesting it's actually some of the early it's actually some of the early employees kind of want to basically out employees kind of want to basically out one of the first of engineers at Claire one of the first of engineers at Claire it was an ex Twitter okay so kind of had it was an ex Twitter okay so kind of had that Scala lineage he brought it here we that Scala lineage he brought it here we run on a Twitter stack finagle all of run on a Twitter stack finagle all of that all right serving us really well that all right serving us really well all right for a financial company all right for a financial company particular like having very very strong particular like having very very strong types making sure that we're not making types making sure that we're not making you know subtle errors that can result you know subtle errors that can result in like tremendous cost to us or our in like tremendous cost to us or our customers mm-hmm and Scala is very customers mm-hmm and Scala is very helpful with that cool so I didn't know helpful with that cool so I didn't know about the finagle foundation because you about the finagle foundation because you know we just ran Scala by the bay ed know we just ran Scala by the bay ed yeah right where Finnegan was in the air yeah right where Finnegan was in the air right and it was the second time we right and it was the second time we actually did finagle at scala by the bay actually did finagle at scala by the bay right Flossie original con and then we right Flossie original con and then we kind of do the whole thing together and kind of do the whole thing together and so one of the reasons we want to bring so one of the reasons we want to bring Finnegan out of Twitter more right and Finnegan out of Twitter more right and want to help them basically to kind of want to help them basically to kind of bring the goodness of it to the world bring the goodness of it to the world and we're getting a lot of value Ataman and we're getting a lot of value Ataman we have a micro service architecture we have a micro service architecture mm-hmm and basic finagle is kind of mm-hmm and basic finagle is kind of infrastructure and the glue for all that infrastructure and the glue for all that all right no this is really cool ah so all right no this is really cool ah so neatly given a lot of ideas right how to neatly given a lot of ideas right how to kind of connect more people to this so kind of connect more people to this so but tell me a little bit about kind of but tell me a little bit about kind of your trajectory in engineering right so your trajectory in engineering right so so I mean you've been identified I know so I mean you've been identified I know identified since they were on Ruby identified since they were on Ruby people they switch to scala right yeah people they switch to scala right yeah so I kind of follow them and and so I kind of follow them and and obviously workdays a one of the biggest obviously workdays a one of the biggest users of scholars they send a lot of users of scholars they send a lot of people to to scale by the bay right so people to to scale by the bay right so and you do like the photos part so can and you do like the photos part so can tell me a little bit you know how kind tell me a little bit you know how kind of your career kind of a line of scala of your career kind of a line of scala what you found useful what you found what you found useful what you found hard what do you want to not repeat a hard what do you want to not repeat a car what they want to do better at Laura car what they want to do better at Laura sure so my first experience with Scala sure so my first experience with Scala was probably in 2009 when I was at a was probably in 2009 when I was at a company called tag tag the time was the company called tag tag the time was the third largest social network in the US third largest social network in the US after first was my space then Facebook after first was my space then Facebook then tagged and one of the other then tagged and one of the other engineers it was into functional engineers it was into functional programming he trot who's trying ups programming he trot who's trying ups golly he rewrote one of our services in golly he rewrote one of our services in Scala go on I was not Johan he was the Scala go on I was not Johan he was the CTO that's right this was one of our CTO that's right this was one of our engineers eat I and Johan the CTO look engineers eat I and Johan the CTO look to the service at hate this is very to the service at hate this is very elegant you know what was several elegant you know what was several hundred lines of code he rewrote in like hundred lines of code he rewrote in like 30 or 40 s it seemed to perform equally 30 or 40 s it seemed to perform equally well I was much easier understand and he well I was much easier understand and he said no we can't ship this mm-hmm like said no we can't ship this mm-hmm like nobody here knows Scala there's no good nobody here knows Scala there's no good tooling support for scholars no idea tooling support for scholars no idea supports like I'm sorry this isn't gonna supports like I'm sorry this isn't gonna work out eat I was disappointed I was work out eat I was disappointed I was slightly terrified of Scala could have slightly terrified of Scala could have it you know kind of a simp similar it you know kind of a simp similar reasons mostly because it seemed reasons mostly because it seemed inaccessible to me so we didn't do inaccessible to me so we didn't do anything with it then fast forward a anything with it then fast forward a couple years I'm at a startup called couple years I'm at a startup called lion side we're facebook social gaming lion side we're facebook social gaming and I was had joined as a very senior and I was had joined as a very senior engineer head of engineering VP madryn engineer head of engineering VP madryn didn't work out I somehow became de didn't work out I somehow became de facto person running engineering over facto person running engineering over there okay and I was an intelligent user there okay and I was an intelligent user and IntelliJ Scala plugin came out yes and IntelliJ Scala plugin came out yes so I said all right let me try this out so I said all right let me try this out and this hadn't just come out I'd been and this hadn't just come out I'd been kind of playing with it a little bit kind of playing with it a little bit some new iteration which came out and we some new iteration which came out and we tried it okay we finally have some likes tried it okay we finally have some likes decent syntax highlighting and a little decent syntax highlighting and a little bit of auto completion actually showing bit of auto completion actually showing you this red squiggly zhonya's know you this red squiggly zhonya's know where you make the errors yes it's all where you make the errors yes it's all right on the right okay look this is right on the right okay look this is actually viable like I think people this actually viable like I think people this gave people it felt that it was enough gave people it felt that it was enough of a guardrail for someone who's coding of a guardrail for someone who's coding a job at a try Scala out and even if a job at a try Scala out and even if they don't know the syntax the IDE would they don't know the syntax the IDE would kind of guide you yes so we tried it out kind of guide you yes so we tried it out and we started migrating some of our and we started migrating some of our Java code into Scala it was nice okay Java code into Scala it was nice okay and it worked okay and it was slow and and it worked okay and it was slow and it was kind of cumbersome but like it it was kind of cumbersome but like it did the job and we were happy with it did the job and we were happy with it fast forward another year and a half I'm fast forward another year and a half I'm now CTO at a start-up Cole identified now CTO at a start-up Cole identified yes yes we're hitting all kinds of scalability we're hitting all kinds of scalability challenges and we got there's got to be challenges and we got there's got to be something to make it run and we were something to make it run and we were using a bunch of ruby on rails which is using a bunch of ruby on rails which is great but like you can't write like great but like you can't write like persistent services that are running and persistent services that are running and processing things in memory like no you processing things in memory like no you got like you rut essentially it's almost got like you rut essentially it's almost like a PHP model rather start a process like a PHP model rather start a process you do stuff and you're done so we we you do stuff and you're done so we we tried Scala for the I knew Scala nobody tried Scala for the I knew Scala nobody else in the company actually did but I else in the company actually did but I kind of encouraged some engineers to kind of encouraged some engineers to pick it up they ran with it we're really pick it up they ran with it we're really successful nice and they did a lot of successful nice and they did a lot of graph processing right so this is kind graph processing right so this is kind of hard to do in Ruby right on the I of hard to do in Ruby right on the I mean tutor switched away from ruben foo mean tutor switched away from ruben foo yeah same reasons yeah so then we yeah same reasons yeah so then we started processing our big data in started processing our big data in scalding mm-hmm Hadoop yes and scalding scalding mm-hmm Hadoop yes and scalding was a really nice way of doing that I'm was a really nice way of doing that I'm not gonna her office or something yeah not gonna her office or something yeah yeah so that was that was really helpful yeah so that was that was really helpful then we get acquired mm-hmm by workday then we get acquired mm-hmm by workday yeah and we kind of became the data yeah and we kind of became the data products team at work day because data products team at work day because data analytics machine learning and we went analytics machine learning and we went from like a half Ruby half Scala shop at from like a half Ruby half Scala shop at that point to one hundred percent or at that point to one hundred percent or at ninety percent Scala whereas all big ninety percent Scala whereas all big data all the time all the services were data all the time all the services were writing in Scala we'd moved up over to writing in Scala we'd moved up over to spark mm-hmm which is all you know Scala spark mm-hmm which is all you know Scala base yeah and I kind of never look back base yeah and I kind of never look back and then incidentally when I was looking and then incidentally when I was looking for transition and I chose Clara as my for transition and I chose Clara as my the next step for my career they were the next step for my career they were already on skull all right so I got all already on skull all right so I got all right all right right yeah no this is right all right right yeah no this is great this is great so Simon still works great this is great so Simon still works for Simon team which was the former for Simon team which was the former identify teams to work this they're very identify teams to work this they're very happy and growing yeah we you know we've happy and growing yeah we you know we've met a lot of folks you know what one of met a lot of folks you know what one of my good friends works there so that's my good friends works there so that's yeah that's that's very good so kind of yeah that's that's very good so kind of Ewing include amount of Scotland startup Ewing include amount of Scotland startup world and kind of the benefits a world and kind of the benefits a comeback I would kind of give you a kind comeback I would kind of give you a kind of an interesting data point and I of an interesting data point and I wonder what you think about it so I wonder what you think about it so I found that in the last year yeah several found that in the last year yeah several fin tech startups wish wish I kind of fin tech startups wish wish I kind of called loosely called loosely protected by anything so basically we protected by anything so basically we should have something do finance start should have something do finance start using scholars so there are several using scholars so there are several companies we should be using Bitcoin for companies we should be using Bitcoin for whatever reason uh you know Bitcoin whatever reason uh you know Bitcoin blockchain actually has a lot of open blockchain actually has a lot of open source projects using scholars own so source projects using scholars own so that that actually is interesting it that that actually is interesting it then there is not a communicant alley then there is not a communicant alley and they recently joined us the funnel and they recently joined us the funnel by shasta which finds you know a lot of by shasta which finds you know a lot of scholar companies right including scholar companies right including website flight band and so they want to website flight band and so they want to build a new bank they basically want to build a new bank they basically want to effectively refine your credit cards effectively refine your credit cards right take a better line of credit to right take a better line of credit to repair credit cards and let you manage repair credit cards and let you manage that mm-hmm and when i talk to them what that mm-hmm and when i talk to them what was the reasons they chose scholar they was the reasons they chose scholar they mentioned several of the measurement mentioned several of the measurement correctness right I got the financial correctness right I got the financial situation but they also mention that it situation but they also mention that it makes it potentially easier ah to enable makes it potentially easier ah to enable acquisition because financial com is no acquisition because financial com is no job a lot of banks are using Java right job a lot of banks are using Java right so basically the the already of GBM and so basically the the already of GBM and so the reason goes that scholar runs on so the reason goes that scholar runs on gdm at the four it's easy to interrupt gdm at the four it's easy to interrupt with larger companies and so I wonder if with larger companies and so I wonder if you know first of all kind of mr. you know first of all kind of mr. present to me right because you think present to me right because you think finance of their conservative industry finance of their conservative industry but here's actually an argument for but here's actually an argument for legacy compatibility 3g am right in in legacy compatibility 3g am right in in kind of a technology which would enable kind of a technology which would enable potential integrations with a bigger potential integrations with a bigger companies i wonder if you guys thought companies i wonder if you guys thought about that or if not you know like what about that or if not you know like what is what do you think about this line of is what do you think about this line of reasoning yeah so in our case we really reasoning yeah so in our case we really really do not want to be acquired mm-hmm really do not want to be acquired mm-hmm so we're it's basically transform so we're it's basically transform consumer finance or go home mm-hmm but consumer finance or go home mm-hmm but the reasoning makes a lot of sense so I the reasoning makes a lot of sense so I know when workday was looking to acquire know when workday was looking to acquire identified they were on the JVM stacked identified they were on the JVM stacked running mostly Java mm-hmm we were running mostly Java mm-hmm we were running a lot of Scala yeah so for them running a lot of Scala yeah so for them there's all this naturally slots in we there's all this naturally slots in we know how to run Java we've got a lot of know how to run Java we've got a lot of JVM expertise so if they needed to JVM expertise so if they needed to transfer people to our teams that's transfer people to our teams that's something they could do so definitely something they could do so definitely played into their calculus for that played into their calculus for that acquisition so I'd make sense that other acquisition so I'd make sense that other FinTech companies if acquisition is FinTech companies if acquisition is a potential outcome they're looking for a potential outcome they're looking for mm-hmm and that's something you should mm-hmm and that's something you should consider yeah and I mean I don't mean consider yeah and I mean I don't mean that note Ali wants to acquire that's that note Ali wants to acquire that's the reason I write like I mean obviously the reason I write like I mean obviously everybody wants to be yeah very big and everybody wants to be yeah very big and I think another question is I'm just I think another question is I'm just curious can feel you know a business curious can feel you know a business model right that like you know to play model right that like you know to play here you lot of capital yes right so I here you lot of capital yes right so I think if you want to be a bank or if think if you want to be a bank or if you're going to facilitate mortgages you're going to facilitate mortgages connect a little bit more how do you add connect a little bit more how do you add value to the horrible mortgage business value to the horrible mortgage business I just I bought the house last year's I just I bought the house last year's like basically I mean to me by the house like basically I mean to me by the house is just my cousin PDF right basically is just my cousin PDF right basically it's PDF and electronic signing and and it's PDF and electronic signing and and ridiculous credit approval through ridiculous credit approval through slow-moving banks right which is so like slow-moving banks right which is so like it's broken many many levels so I wonder it's broken many many levels so I wonder you know how would you want to prove it you know how would you want to prove it with all the technology yeah so it's with all the technology yeah so it's interesting mortgage banking is really interesting mortgage banking is really an information processing problem mm-hmm an information processing problem mm-hmm because what what what does it mean to because what what what does it mean to get a mortgage it's kind of like filing get a mortgage it's kind of like filing your taxes that's kind of filling out your taxes that's kind of filling out the mortgage application yep and then the mortgage application yep and then getting audited at the same time correct getting audited at the same time correct so and that's not very fun yes what so and that's not very fun yes what we're trying to do is streamline and we're trying to do is streamline and automate as much of that as possible so automate as much of that as possible so instead of you getting on the phone and instead of you getting on the phone and dealing with your broker or loan manager dealing with your broker or loan manager whatever they were ever they were and whatever they were ever they were and concentrating so what's going on what's concentrating so what's going on what's what's happening what do I need to do what's happening what do I need to do right maybe that maybe sent me up there right maybe that maybe sent me up there calling you and telling you what you do calling you and telling you what you do instead imagine just an online instead imagine just an online experience mm-hmm where you go experience mm-hmm where you go step-by-step you come well collect step-by-step you come well collect whatever miss you need you put it all in whatever miss you need you put it all in there mm-hmm anything most everything is there mm-hmm anything most everything is automatically verified via API automatically verified via API integrations mm-hmm and there's all integrations mm-hmm and there's all kinds of complexity that today is done kinds of complexity that today is done by you know got processors and by you know got processors and underwriters and compliance people underwriters and compliance people locked desk you know capital markets and locked desk you know capital markets and closers funders I mean there's probably closers funders I mean there's probably like five other functions I'm not like five other functions I'm not mentioning which is incredibly human mentioning which is incredibly human labor intensive yes all of that being labor intensive yes all of that being streamlined orchestrated by either streamlined orchestrated by either workflow systems or whenever possible workflow systems or whenever possible automated automated and as a consumer you can see what's and as a consumer you can see what's happening step-by-step so it's totally happening step-by-step so it's totally transparent to you mm-hmm you know when transparent to you mm-hmm you know when you're more just gonna get close you you're more just gonna get close you know what the bottlenecks are and you know what the bottlenecks are and you always know what's on you mm-hmm so the always know what's on you mm-hmm so the other day what was this horribly painful other day what was this horribly painful process where you have no idea what's process where you have no idea what's going on or even if you're going to get going on or even if you're going to get the mortgage this becomes like open and the mortgage this becomes like open and transparent yes and so heavily automated transparent yes and so heavily automated it probably takes a small fraction of it probably takes a small fraction of the time can you do you plan to give her the time can you do you plan to give her guarantees over closable time window guarantees over closable time window will take eventually yeah we're not will take eventually yeah we're not there yet because that's a thing that's there yet because that's a thing that's very important right in the area right very important right in the area right they demand very short closing windows they demand very short closing windows yeah and then the banks could not yeah and then the banks could not deliver it the rest of the traditional deliver it the rest of the traditional banks a very slow so i'm gonna use banks a very slow so i'm gonna use traditional banks and basically again i traditional banks and basically again i used underlying banks and for financing used underlying banks and for financing or your own so sort of hmm we're not or your own so sort of hmm we're not inventing a new model for mortgages inventing a new model for mortgages we're not creating new sources of we're not creating new sources of capital and some players in space are capital and some players in space are mm-hmm we the way it works is we partner mm-hmm we the way it works is we partner with warehouse banks kind of like with warehouse banks kind of like wholesalers for money yeah so when we wholesalers for money yeah so when we close your loan we have a credit line close your loan we have a credit line with them it would literally just wire with them it would literally just wire transfer the money from our from the transfer the money from our from the warehouse bank to your escrow account warehouse bank to your escrow account and bloom you have your loan you have and bloom you have your loan you have your home and you can move on mmm-hmm we your home and you can move on mmm-hmm we then sell the loan to seller aggregators then sell the loan to seller aggregators mm-hmm these are the these are companies mm-hmm these are the these are companies that purchase these securities these that purchase these securities these loans yes from other mortgage banks loans yes from other mortgage banks right and then they sticking bundle them right and then they sticking bundle them into what this coach or Julie she the into what this coach or Julie she the mortgages what's that that so like this mortgages what's that that so like this Sora composition shade mortgages right Sora composition shade mortgages right now they are something quickly to now they are something quickly to exactly it was aggregators exactly and exactly it was aggregators exactly and eventually will become the aggregate eventually will become the aggregate ourselves as we get big enough sui will ourselves as we get big enough sui will dissing me disintermediate them we've dissing me disintermediate them we've already disintermediated brokers we're already disintermediated brokers we're doing that part like talking to the doing that part like talking to the customer directly ourselves hmm and then customer directly ourselves hmm and then we'll eventually sell directly to the we'll eventually sell directly to the capital markets mm-hmm whether that's capital markets mm-hmm whether that's Fannie Mae or Freddie Mac or other Fannie Mae or Freddie Mac or other investors but we're going to interface investors but we're going to interface directly with them so the goal is to directly with them so the goal is to create a short of a distance as possible create a short of a distance as possible between you know a home buyer consumer between you know a home buyer consumer and the sources of capital nice so you and the sources of capital nice so you know I person found like red so know I person found like red so the problems with banks with mikey is the problems with banks with mikey is right like no let's say an experience is right like no let's say an experience is our own social security number mm-hmm our own social security number mm-hmm for my wife the completely stops bank of for my wife the completely stops bank of america yeah the basically said go back america yeah the basically said go back to experience fix it mm-hmm right and to experience fix it mm-hmm right and then proceed yeah and somebody told me then proceed yeah and somebody told me Bank of America will never close in time Bank of America will never close in time and I said no they give a very good and I said no they give a very good right I'm going on dreamin customer yeah right I'm going on dreamin customer yeah they know they won't do this right they they know they won't do this right they almost cry they asked me to consider the almost cry they asked me to consider the fact that they will not be able to do fact that they will not be able to do this I feel extremely short closing time this I feel extremely short closing time so I just for fun just to entertain the so I just for fun just to entertain the humor this person I initiated the humor this person I initiated the process with two mm-hmm because and I process with two mm-hmm because and I felt bad about this because one of them felt bad about this because one of them is not gonna get the deal right they're is not gonna get the deal right they're gonna work for me and not i'm gonna say gonna work for me and not i'm gonna say no you know i'm with the other one no you know i'm with the other one because you know maybe if there is a because you know maybe if there is a different and bank america failed different and bank america failed because experience have there in a because experience have there in a bureaucratic mistake that's completely bureaucratic mistake that's completely substance on that tracks they say go substance on that tracks they say go back to experian you cannot even call back to experian you cannot even call experience yeah right there's a whole experience yeah right there's a whole bunch of you'll experience a mic reports bunch of you'll experience a mic reports is wrong give me credit I'm mccray score is wrong give me credit I'm mccray score is wrong because of some states right so is wrong because of some states right so so basically there are very fortunately so basically there are very fortunately the second worker was able to do this the second worker was able to do this right so so it was very grateful to the right so so it was very grateful to the person who advised me right so the person who advised me right so the traditional bank apparently post-crisis traditional bank apparently post-crisis right there basically super conservative right there basically super conservative so what are right but you know the so what are right but you know the infrastructure is he right there is infrastructure is he right there is transgenders experience right Equifax transgenders experience right Equifax and so apparently for bike market and so apparently for bike market secrets this regulations mm-hmm you know secrets this regulations mm-hmm you know numbers of the match two out of two much numbers of the match two out of two much third doesn't match go fix it right the third doesn't match go fix it right the basically right so so I were able to to basically right so so I were able to to fix this kind of errors right we should fix this kind of errors right we should throw a monkey wrench into the whole throw a monkey wrench into the whole process are you are able to deal with process are you are able to deal with this all right because if you're one of this all right because if you're one of the banking bank itself right back we'll the banking bank itself right back we'll do the same thing even if you do it do the same thing even if you do it online so so I were able to apply the online so so I were able to apply the intelligence right like an intelligent intelligence right like an intelligent person looking at this probably will say person looking at this probably will say this should not stop right my this should not stop right my application mm-hmm so how can I deal application mm-hmm so how can I deal with this kind of stuff ish current with this kind of stuff ish current people perceive as kind of bureaucratic people perceive as kind of bureaucratic nonsense that's a great question so one nonsense that's a great question so one of the things that we're trying to do of the things that we're trying to do with our workflow systems our with our workflow systems our automations is actually leverage people automations is actually leverage people where people where people or what people are best and that's or what people are best and that's applying judgment mm-hmm so whether it's applying judgment mm-hmm so whether it's this case or some other like non happy this case or some other like non happy path case mmm that maybe the automated path case mmm that maybe the automated systems or the workflow systems can't systems or the workflow systems can't handle mm-hmm we head will have an handle mm-hmm we head will have an exception flow mm-hmm where computers exception flow mm-hmm where computers handle the easy stuff and if something's handle the easy stuff and if something's out of the ordinary it goes to people out of the ordinary it goes to people mm-hmm whose job isn't to make a routine mm-hmm whose job isn't to make a routine decision but actually to exercise decision but actually to exercise judgment mm-hmm which in this case would judgment mm-hmm which in this case would be a perfect example of okay like be a perfect example of okay like obviously this is okay there's just like obviously this is okay there's just like one out of three is off let's just one out of three is off let's just ignore them and move on yeah and it's ignore them and move on yeah and it's not the primary right so it's kind of not the primary right so it's kind of so-so but that will be possible because so-so but that will be possible because you will have your own judgment about you will have your own judgment about this warehouse bank credit yeah well I this warehouse bank credit yeah well I think there are certain restrictions think there are certain restrictions that's like what Fannie and Freddie Mac that's like what Fannie and Freddie Mac are willing to deal with mm-hmm but are willing to deal with mm-hmm but that's not that's not one of them so if that's not that's not one of them so if it's purely internal restriction mm-hmm it's purely internal restriction mm-hmm then we're actually setting people up to then we're actually setting people up to exercise judgment rather than setting exercise judgment rather than setting people up to follow a strict process hmm people up to follow a strict process hmm I think it's kind of just in terms of I think it's kind of just in terms of how or how we're organizing ourselves how or how we're organizing ourselves that's quite a big this so it looks like that's quite a big this so it looks like you should be a mortgage company it you should be a mortgage company it sounds like you have a lot of machine sounds like you have a lot of machine learning to do right because some of learning to do right because some of these things I can be scary yeah so these things I can be scary yeah so whatever plans from ml and machine whatever plans from ml and machine learning and artificial intelligence you learning and artificial intelligence you know can you become a guy company know can you become a guy company mortgage space interesting so we are mortgage space interesting so we are going to apply a I it's going to be going to apply a I it's going to be around a few different things it's going around a few different things it's going to be around customer lifetime value to be around customer lifetime value mm-hmm try to unfinished I what kinds of mm-hmm try to unfinished I what kinds of people should we have put more emphasis people should we have put more emphasis on it's going to be risk analytics hmm on it's going to be risk analytics hmm trying to gauge what is the credit risk trying to gauge what is the credit risk that any given person poses that's that any given person poses that's actually going to be very valuable to us actually going to be very valuable to us if we can demonstrate a superior risk if we can demonstrate a superior risk model based on AI we can actually market model based on AI we can actually market our mortgages to investors for a bigger our mortgages to investors for a bigger premium you literally have higher premium you literally have higher revenues if we're really good at AI revenues if we're really good at AI mm-hmm so that that's probably the most mm-hmm so that that's probably the most immediate application interesting and immediate application interesting and how far are you from this that we're how far are you from this that we're still a ways we're probably going to still a ways we're probably going to start doing our first predictive start doing our first predictive analytics more on the user acquisition analytics more on the user acquisition side like more marketing side of things side like more marketing side of things and maybe q2 q3 of next year okay and and maybe q2 q3 of next year okay and not too far away not too far away not too far away not too far away lifetime and start up yours yes yes yes lifetime and start up yours yes yes yes yeah and then more on the risk modeling yeah and then more on the risk modeling probably late next year early the probably late next year early the following year okay and do have any idea following year okay and do have any idea which sticks which software sex again which sticks which software sex again use for Marceline not yet okay not yet use for Marceline not yet okay not yet all right captain water down adopts all right captain water down adopts we're gonna write a post one of them we we're gonna write a post one of them we will cover by area in addition to the as will cover by area in addition to the as a scholar and the some spark right now a scholar and the some spark right now so yeah so I'm a lib and sparking I'll so yeah so I'm a lib and sparking I'll you know it's all kind of good options you know it's all kind of good options at this point yeah cool so super at this point yeah cool so super exciting looking forward to your talk exciting looking forward to your talk about chloride or when meet up and thank about chloride or when meet up and thank you very much for foreign skier all you very much for foreign skier all right thank you Alexi what