Distributed Harmony: Unlocking Agile Delivery in a Multi-Source World - Sasha Pitkevich | EPAM

Hi guys, I hope you would be able to see my presentation soon. Uh I'm Sasha Page. I'm leading blockchain professionals community and for last 10 years I've been leading blockchain practices at big enterprise system integrators. Uh but what I'm going to speak today is about the evolution of the project management. Basically we all witnessed uh for last five years the situation where like there are a lot of questions where the blockchain technology is applicable and we all know that the blockchain technology is applicable the best and brings the best value where that's the business between legal entities where we are either combining in a one solution several legal entities either we are combining in one solution uh individuals where that's a public solution In both cases, that's the business in between businesses and current economical and political situation requires us to develop such a solutions on a distributed way.

when that's the public solution which needs to be uh developed that's normally developed by the community and we have a need to evolve this solution by the community and even when you start as a small group or as a startup trying to scale this solution you need a community to help you and when we are speaking about enterprise solutions we normally have several probably competing organizations who are stuck in building the software altogether who are stuck in managing this project on the way this project should evolve and uh looking to the current economical situation where uh that's not anymore where we just give money and we are not counting them. The current state of the market doesn't allow that even to the big enterprises. They started to count every penny which means that they started to be particular about the project management and how the money is spent which means that they count money where uh the project is being developed and combining development with other legal entities. They started to divide the money and they started to divide the power and all of this needs to be managed and a majority of the traditional project management even like what the most efficient and most well known now is agile and normally this all like uh when we are uh in a public space when we are developing with the community we are trying to implement it somehow and the Open source development is is is a good uh example of this. Well, we are doing that in enterprise tech.

We have the more uh more managed agile approach where we have teams and a set of managers on top of that making the delivery predictable. In both cases, the desire is to make it predictable and to lower the cost. uh in both cases uh the model of the open-source software development looks attractive but when we look to the reality I mean I don't have I do not see because of the light of many people and uh my idea was please raise a hand those who participated in a open-source software development for the majority of the cases that's complex and annoying because you need to agree with other people who are doing something and it's very very difficult to coordinate where you develop the same way in enterprise environment that's even more annoying because you have a set of managers and those are your managers not your managers and they all trying to coordinate all of this in both cases that's happening kind of mess and that's why we started to think about the new algorithms of how to manage the software delivery Okay, which will efficiently satisfy the need of the enterprise software development and public software development where we end up with the good quality of the product and we know which money how much money we spend on that and we know by what time it's being delivered. And looking back to the traditional open-source models, we started to think like what is important? First of all, the team and the most important who creates the code.

Code is being created by developers. If according to agile manifesto, we have experienced developers, they normally are able to manage themselves. if they able to manage themselves and if they have the media to talk to each other. Do we really need the set of the managers on top of them and how much this managers can add to that and again we all know that the blockchain solutions are meant to automate the managerial decisions. So we are coming to the most logical thing ever is why not to apply the algorithms of the traditional to blockchain to the management.

Why cannot we try to automate the managerial decisions during the project management and here we are not uh like we are not inventing something totally new. You all know like I I guess I'm talking to the professionals. You all know the abbreviation of DAO decentralized autonomous organizations. And those bring us one of the implementations. There are many but that brings us one of the implementations of the decentralized governance model.

And here basically we're speaking of various ways and there are various mechanisms which lays in a DAO implementations and uh you know that basically there's a managerial approach of the decentralized governance which normally based on the communications of gathering the signals and making the decision based on the signals with the either majority either erized majority here like again there are many algorithms how we make a decision. The important thing is that those decisions are made on a consistent way and uh the signals and the decisions are recorded and recorded on a best like in mutual ledger and uh that's the algorithm of decision making is predictable. Yes, that's that's a smart contracts we're saying. smart contracts which are laying in a in a DA model and uh several models leads us to the theory it can be done well then we are looking to the practice and and I will give the examples in a minute like basically we are looking to the uh examples of how it's happening and again majority of the audience now looking to the open source projects you you participated with and saying like okay that's never was uh where successful because in the majority you are stuck in a slow decision making because well normally you have to get the signals, you have to analyze the signals and you have to somehow facilitate the decision making uh which is normally like voting of the participants and uh all those discussions being not put in a computer being not put in a smart contract they takes time they consume time and slow down the project and also the documentation of those decisions and a history of decision making normally is not kept well intact and all of that causes the complaints and it causes to uh like basically the mess for the people who are participating and we're ending up of the professionals being theoretically able to deliver the good code but not being able to work efficiently together. And here again we see the clear motivation to put that in a computer because we don't need to be stuck of running among people and saying uh come on you still didn't give your vote or come on we are waiting for you or are you sure or what did we decide two months ago and uh this all leads to the try to implement the contracts around this And we do see several models already being implemented.

We have what the simple model talking weighted voting. That's what uh ENS DAO used uh to manage their decisions and they were quite successful there. Uh quadratic voting much more complex. I wouldn't recommend this algorithm for those who are just starting with the model. But this all you know you're choosing the DAO governance uh basically tied to the size of your DAO to the experience of your DAO to the diversity and how they are familiar with all the crypto mechanisms of because you're not necessarily speaking for the people who are native to crypto.

The same mechanisms are applicable for the project management in a projects which are industry applicable where crypto is not a core of their solution where they using blockchain technology for example to do the traceability of something. Uh so quadratic voting much more complex but used by city DA and used quite efficiently there and reputation based voting those you all know maker DA uses that for many years already right and multi- signature governance another way which is uh very suitable for not that experienced in crypto people and also helps to distribute the decision making where again remember the enterprise case where you have teams owned by the various companies who are sometimes competitors and still need to develop the same set of the software altogether or even to coordinate the development of the competing pieces of the software uh but on a coordinated way because they underly on the same uh level one network and for them the multi- signature governance can simplify significantly can be used and that's I think it's used in the safe DO yes and they and there are also like the good thing here is is a smart contract so basically how the act your DAO is being managed and what are the rules you can program and you can define and you tailor that for the particular need as I already said depend on your dow size speed of the evolution which is impossible communic communications uh which are existing inclusivity because why and in the desired level of the decentralization because that's also you manage that and the hybrid versions for example in the central end they all they use the the hybrid level of governance And what I'm saying is the current market it really asks for those ways for those possibilities to manage our projects cheaper and our projects require in many cases decentralized governance and decentralized governance is the best way implemented on chain. It gives us more control, more flexibility, more transparency and we already have this like but governance model and the D model it's just the start while we are doing the project management we need other things to be implemented and here we're speaking about the difficult and change management all the delivery projects they have backs they have changes and normally the endless pock of the changes which which needs to be managed and here probably for you that's native you know about improvement proposals concept and you know about bounty concepts two weeks ago I was presenting the same pitch for like much more enterprise agile and scale conference and for people I was not that native all those approaches but they resolve the need of the decentralized defect management where we can vote for the severity and the priority of the defect resolution or where we can vote with the money for the certain changes being implemented which brings actually very native and market native mechanism for your solution to evolve because the solution started to add the functionality not the one which is desired by someone who has his personal vision by But by the community decision where in many cases the users of the solution vote with the real money for what needs to be implemented which tights your solution much more to the reality than where the individual business is doing so. And uh last but not least, while we are doing the centralized development with the distributed and decentralized governance, there are always the questions of access and code ownership. And for those uh there are models of the open source, the hybrid model and the closed source.

You can decide as a team or as a community which part of your code is uh public or is uh applicable for the changes. you can still keep the core of your solution if that's the business decision close and still allow the far community uh to have the outskirts of your functionality outskirts of your API or something to evolve being communityled that's the another degree of flexibility you can apply here and uh there are already the mechanisms and like I'm here a little bit of discussing the future There are mechanisms of the code tekkenization and the next example is well I will probably move to this slide because that gives you the real mechanisms. I don't know how many people here heard about the crypto pond. If not I'm really advising you to go there and check because that's the mechanism to get onchain data for uh large language models. That's the mechanism where we can use onchain data for our LLM analysis and for the building models which helps and gives you boost for your DAO and LLM being combined but also gives you the boost uh to analyze the involvement of the code because what when like I'm here in a more public like less enterprises atmosphere so you guys are mostly thinking about how to how eist system invests in the various projects.

Here if you will go to the crypto pond you can see that there are different models there and one of the recent competitions which is still ongoing and which is uh initiated by vitalum is about analyzing GitHub sources uh in terms of the uh involvement or in terms of the implications on a ecosystem. Basically now the teams are competing and we uh as a company also creating our models because it gives you instrument of how to analyze the participating projects in ecosystem. uh in terms of the fair reward in terms it tells you educated like it motivates the educated decisions based on the previous history and and the potential uh where to invest and that's like why I started to speak about the code ownership and uh monetization of your development uh involvement. This is a way how we can resolve that signal based like information based where we can make grounded decisions. Those who ever participated on application for Ethereum grants know how messy and unclear the system is.

And that's the way out of this and that's the way which you can use on your projects within your ecosystem and which enterprises can use in terms of analysis of who participated the most for the solution even when the ecosystem grows enormously big. So please don't forget about this and uh another part I I almost cover it like management decision making managerial decisions code ownership defect management change management almost everything we need for the distributed centralized project management. The last one is a community management. this the last one not least one other people because we need to keep our community engaged and again all the mechanisms I've been speaking about they help you to tight the decision making distribution of the power and incentivation monetization to the actual participation and enterprise cases we are speaking about the corporate participation in the public cases we are speaking about the individual particip ipation and here what besides the fact uh all the communications all the community all the social stuff needs to be taken care of properly because when we're speaking about the like decentralized project management we might like what for many people who are uh less public in the more enterprises they think okay I'm putting everything in the code and I'm done sorry it's Not because you still have to take care of your people. You still have to take care about the information flows in between people because that's appear to be the key if we are gathering signals and again the uh loyalty of the people their motivation that's not always and we hear that from many HRs it's not always this in a in a monetization space not at all and we know all of this because of like we are here on Ethereum community conference and you know that we have diff almost every year because we need to keep the level of knowledge we need to keep the level of the engagement we need to hear more from a community and that's that's truth for any size of the communities so by like uh I spent quite some time explaining you how to put everything in the computer but here I'm devoting time to say yes you put the governance in a computer and in a contract in a down mechanism but relationships in between people needs to be taken care of and that means that probably instead of the project manager who is making the decisions individually we will we will shift this power to the DAO mechanisms but we will need a community manager who takes care about the well-being of the people and being well informed on a project.

And uh I already spoke about how to get the analytics and uh initially when I was speaking on this slide and I was speak I was more speaking about the analysis of the source code and uh uh and who does what and how much is is the involvement to tie that to the monetization but at the same time Yeah. But at the same time uh we are using the same mechanism for do the project analytics because if we have the governance on chain we get the information on chain and by the way crypto pond gives it to you for free which means that you have implement the feedback analysis mechanisms using onchain data on chain governance data if you put it there and uh and then have like you know the feedback loop on your management and I'm at the end of my presentation and what I tried to tell to you is that it's exactly the time when we can think about the future project management being done by machine. We must not forget about the real souls of the real people participating but the pure management tasks can be done on a distributed basis and can be done fairly. questions guys it's not only future that's also reality I gave you the examples there are also tools like noises like which set up you like which helps you to set up that and to implement that so don't think that it's like in 10 years it's now and you can try that already now thanks