Add Observer Signature with airSlate SignNow

Add observer signature

hi everyone thanks a lot to come therefore I will need to give this back to you so Who am I to come here today to speak to you the city of a company called quick sign we do digital onboarding I've been in this company since five years before that I was consistent for nine years in a French conceptive company and I've been designing enterprise architecture for quite some time and challenge quiz I'm to help them in everything related a business query so pizza is one of the European leaders are in digital boarding for financial services we now handle a millions of financial transactions per year which wouldn't see them wouldn't seem that much compared to those industries but in terms of financial transactions it's actually quite a lot we are not very well known from the general public because our assertion is actually used under the hood by big companies like banks insurances that we are always integrated in white labelled so you as a customer we never see our - so which is one of the reason why I am here today it's to give a bit visibility to what we do at quick sign to engineers and you as an end user we provide care I see 4cr and Electra and if sonic signature features with customized workflows and our objective is to drive back refuse integration using OCR and new virus operational efficiency in insurances and banks during the end-user onboarding process because onboarding new customers is obviously very important for a lot of industries so the process is a represent to optimize we've been doing that for 10 years we are in 15 countries and we have a kind of kind of a good growth rate in terms of the number of transactions that we handle on the platform each year in the 70% more than the year before so are the pressure our platform is growing quite a bit so as I said we are used by many near banks and insurance companies and also retail industries those are one of the key customers that use our platform we have BNP HDC engineer range bank for one of the latest one so well we know but we are actually used by important industries why change anything I mean we've been doing this for 10 years it's working good our growth is sustained so why would we want to change anything well our platform has beaten - as getting too hard to monitor it's too heavy it works but is being built around in-house workflow by configuration engine I mean like many companies in our field I guess and it has gotten to the point where changing anything is too complicated it's a memory so we had gotten to the point where we had to change something it's too hard to scale it wasn't designed for that at the beginning it wasn't designed to be messy talents so I think with a tendency to a platform which wasn't built for that since the beginning is quite a challenge well and we took too much time on fixing defects and was first rating so we had to change this what did we need well the first time to market I mean who does I want that I think everything in this room wants that to ship for that faster so that's no new news scale in practical terms but also in business terms because ok you we are a sustained growing rate so we will be able to have more transactions on our platform but we also want to scale the business from a business standpoint what what does that mean on the markets in which we are currently already implemented such as France Germany Italy well okay servers are ever-changing needs they want to optimize their processes the war they want more efficiency and rightfully so that's why the the user platform but we also want to get into new countries new local markets where the rows are different and we need go to adapt to local legislation which is very different from country to another little bit faster and more features I mean we have Atticus that is a data science team we do all the latest fancy stuff and very nice new features but it isn't it isn't useful if you cannot get your admitted SMS feature into your product so getting integrating those features into the product was very important for us as a ten indication of how hard it is the very first data science feature that we developed in our lab took one year to end up in our product it's it required it was before we changed the platform today we don't know one year I mean we want to wish and with a new platform it's possible even this more agility we've secured the problem and methods even using agile methodologies such as scrum generations but it's useless to do I mean it's my point of view and not everyone has that one of you but my boy who is at issue you can be as agile as you want from a methodological point of view if you are driving a tanker you will know the faster you will not turn faster so you need to have to be John at the core technical technological level so yeah we've submitted a very sort of steps such as auditability multi-tenancy security and resiliency by design and cut operational cost and complexity I mean our premise platform is kind firstly complicated to offer it and we did not wander and want that anymore and trying to work with I mean we're a tech company we are reverse the engineers the product that we develop we work on it every day and we want to be able to have fun working on it and not suffer so we decided to go be parent centric buy and hold our features without them being orchestrated by a bit of an engine such as smart document management I mean we are in the digital subscription business so one of the core feature that we have is collecting your personal information and your personal documents so that's one important area now electronic signature we sign contracts trust ability we're in a very somebody gave me the word this morning during the lunch and not forget it but our compliance thank you we're in very compliancy already markets so we need your food thank you traceability yeah more right I mean where we are doing a lot of AI oriented features statistics and kyc know your customer whatever also micro services very equipment will get getting agility so what we do I've talked about the mitigations now how did you tackle the problem with what's going to tap out secures yes architecture that's kind of complicated training but I won't go too much I tried to keep it as an overview but enough so that together you get the point so we have a REST API one of the key things that is that we separate the rights for the reads and on the right side we send we call Canada I mean using it's out of the box whereas the u.s. API and the calendar engine depending of the person definition that is the price calls a various of our features such as our ER engine or fraud detection engine or our signature engine or third-party services because we also integrate a lot of third-party services all those where services whether there are internally developed at quick side or externally available are now integrated in a platform as micro services and they can also they can be either synchronous when you call them they give a reply which is a bit different from what was mentioned earlier with external tasks or they can be a synchronous and in Canada you can implement as synchronously either with external tells but there are various ways eczema tasks are one that we didn't build that way for health reasons we instead chose abuse oriented approach using Kafka because we need food traceability so I am sent for many use cases tasks exam tests are a much easier and you should go for it but we needed to actually trace everything going on on our platform which is why we chose to go through Kafka Elam here comes the interesting part which is kind of posture on on our architecture is that we don't go through calendar for reads we go through calendar only for writes for the rich we use leverage the history has bi which is available out of the box in calendar and we have done our own implementation of that SPI which sends each and every tiny BPMN events to a Kefka topic which we call internally the history streamer it's an append-only immutable logs which means that you never modify it you only add to it and you can consume it from as many microservices as you want and as many times as you want which is extreme flexibility and a lot of very interesting properties we consume I mean it's one of the backbones of our architecture we consume that history stream for many micro census but I don't go into every one of them obviously the two main ones are the case micro services and the search micro service the search is kind of common nowadays we consume those events and we index them in a sick search in order to have full-text search features there is in the camera community a plugin which does exactly that but requires to be deployed inside Karuna which we didn't want we wanted the calendar engine to be the calendar micro service as well it as possible and to me who the weight was alert to other microservices that we could scale indefinitely and as we needed anyway twelve that history stream for trust ability and to add as many micro sizes as we wanted we could as well index from the center page the other master micro service which is original is the case micro service is the one who has a responsibility to create a JSON representation of our case we read all those ten villains and we create a business-oriented JSON representation who contains the states and the actions calendar has such a feature out of the box but the calendar feature is for internal use I mean it's complete you can do whatever you want interact in any way want with the camera engine but we needed our own business oriented representation of the cases are we going to some details after one after and develops the binary contents we don't store them either in Canada or in kept in Casco you should not do that absolutely not so at the entry point of our platform we touch the strengths we can see HTTP while in the HTTP requests we catch all the content which is binary usually it's multi-part but not only and we intercept it and send it to abrazo and instead of sending blood to calendar or calf:cow we send URLs I mean it's very obvious but it's one so something you have to have in mind in the Microsoft platform bridging bridging the gap I mean how do you I mean we had a platform before I mean the the platform in our token are you about today is not yet in production it's going to be a few Montes so we are very close to the end but the previous platform is still currently running and we are we will not do a full replace I mean the new platform is going to augment the previous one so we need a way to integrate them in a similar fashion for that we are left with is the choice of figures back to integrate the situation Sparta pattern inside our previous manner that of our current Minami it was very painful back at the time we did it for just a bit the reasons for compliance reasons but when we started this project we were very happy to add them that show before because we can just take all the events produced by your mother it put them in a casket avec and inject them in commander so that Canada acts as kind of synchronization mechanism which allows us to augment the capacities of our service platform with our new platform and our customers can just use since in both cases both our API are rested us appears and we try to have a compatible educating we can use both our and they get CPI and our new API same as me and benefit from new features and old features and obviously we can come back our previous system using our previous legacy were sit here a relatively generated REST API crazy I don't mean it so it hello something very interesting what a character customer our our API is generated tailored for each customer depending on the specific business process definition of one particular customer with only the relevant actions it can be synchronous or asynchronous some of our more tech where customers want asking Felicity and are able to use a push model using SS SS C or WebSockets but some of our other customers don't want that I mean they also think simpler with using a simple synchronous API so we are able to do both its Delta malware which means that the recipient reflects the the schema of the data which is manipulated by our water engine it has a bug role based access control facility which means we can control what variable and are visible or honored depending on the role of the users or the channel it is multifaceted which mean that on the same process instance we can mix we cannot simultaneously a mobile application and esterification connection and the different face a different view on the sample on the same process instance the other aspect is performance read performance it shows each calendar I mean it's not mentioned this morning Canada is awesome but is built on a database in genesee a related architecture which has its own limits I mean it's blocking and it limits the scale that the throughput that you can have this is fine for us for the right with the amount of Rights that we have on our platform is okay with that limitation but as far reads are concerned it was a limitation to the problem that we needed to address so the rich side of architecture is very fast I mean it's hard to get faster than that and we can scale the read and write separately depending on our end it's pretty good freedom working on the rails I mean the rights are blocking because JDBC in Java is blocking and there is no way around that but under it we are not working a synchronous and the captain's everything almost is in Kafka commands the replies whether they're single or multiple the events such as the calendar history well only for now between tend to be able to generate more events because for now we are a bit around this calendar and calf:cow architecture mainly our case service but Winton's with all our future my processes around it so the our future processes will have to be able to generate events and we want them to be in Kefka and for Canada to go to listen to them and react to them first we for better accessibility we intend to trace where is also if gathered we are not doing it right now yet we wait to be sure to be more confortable with the local tourism before adding the horizon circuit so for the Paris for now we are we are using standard opposite directions routine we wanted to be able to take over the world ranging from the activities from the metro services implementations I have five minutes so instead of counting we wanted to be able to change protocol if we wanted to to change protocol data format going from JSON to or changing from Jason Favreau with our visitors are using cash cow or using standard RPC or rest and without redeploying are off engine so we developed our own service discovery on top of humanities metadata and Carol and we are doing photo colonization content negotiation and exchange pattern negotiation X is synchronous or asynchronous one reply versus materialized it's a truck it depends on tomato service and that single session is done that service discovery channel and semantic versioning routing muscle so we have absolutely no activity specific code deployed inside the camera that micro-sized every single discovered we might get it open social death is there is interest because I mean it solved a problem for us but we are not sure that the program is actually of significant enough interest for everybody for us to do the investment of putting it opposition but if there isn't just where shall we do it where cavenda it's in the works it has very nice for business be German and German editors I mean it's one of the first that when I try I was feeling like ok I could actually use it myself which is not the case for many of the editors of their own it's customizable one of our we wanted an editor a mother which for which we could customize and adapt to our own needs and we have not done it yet but we have posted and we will do it in the near future it has a robust became an engine I mean it works it's easy to integrate it as a spring potato system and we are a Java shop and we love spring so integrating Camilla was very easy for us and very natural the code is it but it's mostly easy to develop except for the core process vector machine which we have put our nose in because we wanted to be sure that we could be able to debug it and fix it in case of problems the core is called hub I mean that's what's next look test roguishness performance because as I said earlier we are not yet some fraction we will soon be but we want to make sure that it's it it's performant and it's robust production in September October around fingers first such improvements study can be optimized [Music] the observer's product is certainly be very tricky for us once we're in traditional ones we have to optimize our existing workflows it's our priority right now was to have something running impression and which is brothers not to optimize our process right away so we wish to me that our GRDC supports migrate to supervisors details but migrants which ability CR is if there are some she balances were people in the room then we were talking about customized designers who are specific needs with our changes with Kissin pellets at a graphical API aside our rest a new CPI and migrate to C be when the right side of architecture becomes a bottleneck honestly I don't know exactly when it's gonna get to be able to like but we know that whether it becomes one we can