 Well, and welcome to this CUBE conversation here in Palo Alto, California. I'm John Furrier, host of theCUBE. We've got a great, awesome conversation with the CEO and co-founder of Harness, a hot startup, Joe Tipensal, who is the co-founder and CEO, but also the co-founder of Unusual Ventures, which is a really awesome venture capital firm doing some great work investments, but also they have great content over there for entrepreneurs and for people in the community. And of course, he's also the founder of Big Labs, his playground, if you will, building out new applications, also well known for being the founder of AppDynamics, a super successful billion dollar exit as a startup, sold to Cisco, and now doing a lot of things and driving Harness, solving big problems. So Joe T, mouthful intro there, you've done a lot. Congratulations on your amazing entrepreneurial career and now your next opportunities Harness, among other things, so congratulations. Thank you for coming on. Thank you, John, and glad to be here. You guys are solving a big problem in software delivery, obviously software is changing the world. You're seeing open source projects increasing in the order of magnitude, enterprises jumping on open source, in general, adoption large scale with cloud, software is being delivered faster than ever before and with cloud scale and now Edge, there's huge challenges around how software is deployed, managed, maintained. You got, we're talking about space too. How do you do break, fix and space? All these things are happening at a massive scale across the world. You are solving a big problem. So take a minute to explain what Harness is doing, why you guys exist, why you jumping in into this venture. Sure, yeah, what Harness mission is to simplify software delivery and make it top notch for everyone. Like if you look at the likes of Google and of Facebook and Netflix and Amazon, these companies have mastered the process of software delivery and your engineers write code and the code is shipped to the end users and they can do it multiple times a day at their scale and at the complexity that they have. But most other business in the world, they all want to be software companies, but it's extremely, extremely hard for them to get there. And I saw this firsthand when I was at AppDynamics as CEO last year, we're about 1,300 employees in the company and we had about 350 or so engineers in the company. And for every 10 or 12 engineers, we had one person whose job was to write automation and scripting and tooling for try to ship software, all kind of scripting kind of stuff. We'll write script in Chef and Puppet in Ansible and to deploy in AWS and whatnot. And one day we were doing the math. We were like, we have about, overall about 30 people whose job was to do DevOps engineering by writing automation, et cetera, to deploy software. And I would do the math like one engineer costs us 200K loaded costs at six million a year that we're spending six million a year just writing, deployment, scripting. And even with that, we were nowhere close to world-class, like world-class as in like what you would think you could ship every day, you could ship on demand, you could deploy software, ship software, all of that. And that was the, I looked at that as a problem inside AppDynamics. And all the AppDynamics customers I would talk to, like large banks and insurance companies and retailers and telcos, and I would hear the same challenge. Like we hear about DevOps, we go to all these DevOps conferences and events and we see the same 10 companies presenting how they home-grew some kind of a DevOps system for software delivery, et cetera. And to me, that was like, we just, we cannot survive with this. Like as a world, we need to have the right kind of platforms for software delivery and simplify this so that everyone could become as good as in Google, Netflix, Amazon, et cetera. That's 10 of our mission at Harness, that can we take every business in the world? And in a few weeks or a few months, can we get them as sophisticated and good in terms of their tooling for software delivery as a Google, Facebook, Amazon, those kind of companies would be. And that's what we are trying to solve. It's a great ambition. And by the way, it's a bold move and it's needed. I'll tell you, it's interesting you mentioned some of those commentary about shipping code at that speed, Facebook, Google, they had to, they had, they were forced to do that. And again, they have all that benefit of the mainstream enterprise doesn't. But if you even go back 20 years ago, 15 years ago, that's when Amazon was born and you see two and S3 is celebrating their 15th birthday. Software, yeah, hyperscalers had some good moves there, but the average business went from craft, waterfall, QA department, go back a little bit slower, I won't say slow motion, but manageable. Now at the speed of shipping and the speed of the scale, that's a huge issue. What kind of pressure do you see that putting on the developer, the individual, not just the system, because you got the system of development and the developers themselves? I think the developers have done quite well to this, I feel. Like, if you look at the software development part itself, the agile development has been happening for quite some time. So developers have learned how to ship things fast and like in a week's print or a two week's print or in kind of faster cycles, they have moved off from the waterfall kind of models like many years ago now. So that's a software and development side of things. Then you have the infrastructure side of things, like can you provision infrastructure fast? Can you get hardware fast? That's the, the cloud has done that well. Where the challenge is the process? The developers are writing code fast enough these days and you have the infrastructure itself could be provisioned and maintained and changed fast enough. But how do you bring it all together and there is the entire process around it? That's not moving fast enough. So that's where the bottleneck is. So I feel the, and the process is not good. The developer experience becomes really bad because developers are waiting for the process to go and they write some code and the code is sitting on the shelf and they are waiting for things. They get both pissed off and mad. What's the hold up? What's the process? And then security shifting left. Wait a minute, I got to go back and rewrite code. This is huge. I want to just get back and just nail a little quickly if you don't mind honing in on the value proposition. What is the harness value proposition? What is the pitch? What are you offering? What are you solving? Can you nail in on that real quick? Sure. So what harness is solving is simplifying the software delivery pipeline. So a developer writes code and then code goes through a bunch of steps. A bunch of steps which is, you build the code, then you test the code, then you do integration test, then you go through your security checks, then you go through your compliance checks, then you go through more testing, then you deploy in a staging environment, then you go do a bunch of things on it, then you start deploying in production environment, but in production you will deploy on like a small part of production, verify everything is working well. If it's not working well, you'll roll it back. If it's working well, then you deploy to more things. This entire process could take like weeks for people to do. And this is mostly automated, you know, in kind of a, you know, this kind of random scripts here and there, et cetera, right? So we simplify the entire process that you could describe your process in the language I just described, like in a very descriptive kind of way. Like this is the process I want to achieve and harness will automatically create your pipelines for this kind of process. And most of these pipelines have a lot of heavy use of intelligence and ML to it, it go from one step to another. Like so many times, like when you say, you know, deploy the code in 1% of my production environment and see if everything is working well. And if everything is working well, go to the next 10%. But how do you figure out if everything is working well? And that's where the intelligence and ML comes in, like, you know, where we learn what is the normal behavior of your application? How does the normal part of the code works? Like, you know, there, what's the performance behavior? What is the functional behavior? What errors it is? And if everything is good, then you go to the next step. So that entire cycle harness automatically, you know, manages and it's automated, you know, if you get governance, you get like, you know, high degree of automation, you get high degree of, you know, security, you get high degree of, like, you know, you know, quality around it. And so it's think of like the CI CD as a lot of developers know this process as is CI CD on steroids, you know, available to you right away. So you sounds like you're making it easier on the CI CD pipelining process. Standing it up, detecting it, prototyping it, if you will, lack of a better description, get used to the pipeline and then move it out, roll it out and build your own in a way. So help them. Is that what, is that what you're doing? Yes, like, you know, a lot of these complex CI CD pipelines, what people need. You know, it can take them like three months, six months to put it, you know, put it together. With harness, it's like an hour, an hour you could put it together, you know, a very, very sophisticated CI CD pipeline. And the pipeline is, you know, automated is, you know, it's intelligent around like, you know, what is the normal behavior of your applications? It's just so fundamentally different than how people have done CI CD before, that we simplify the process, automate the process, you know, and make it manageable and very, very sophisticated environments. And it's funny you mentioned the three weeks that the weeks it could take to do the CI CD pipeline. Of course, that doesn't factor in the, what happens when you roll it out, people start complaining, playing with it, breaking it, then you got to go back and do it again. I mean, that's real. I mean, that's a real problem. I mean, can you just kind of give a taste of the scar tissue that goes on there? What's some of the, what are some of the pain points that you solve? Yes, so I think that is, that really becomes the core of the pain point, people need like high amount of debugability, easy to change things. We call it like the lack of intelligent automation and this heavy amount of developer toil, that the developers have to do so much work around making all of this work. It has to be simplified. So that's where our value prop comes in. You can get like a visual builder and like minutes, you can build out the entire process, which is your sophisticated CI CD pipeline, or you could also do like a declarative YAML interface and just like, you know, in a few lines, just write up whatever process you would want. And we would, we ship with all kinds of integrations with every cloud environment, every monitoring system, every build system, every kind of a testing process, every kind of security scanning. So you can just drag and drop and in minutes you're up and running. It just creates so much velocity in this entire process and also this manageability that people are struggling with. Yeah, morale too. I mean, you can imagine the morale of developers go up significantly when you start seeing that. The developer productivity has always been a big thing, but this intelligent automation conversation is huge. Some people have it, some people don't. Some people say they have it. What is, how can the company figure out if someone's really got the real deal when it comes to intelligent automation? Because again, automation is key into DevOps. Yeah, I agree. I think, I almost call it like, if you look at the generational evolution of things, like the first generation was that developer writes code and then you will give it to some IT admin who will go and deploy the code, run some commands and do things. Generation two was writing scripts that you will write a lot of scripts. That was automation, but it was kind of dumb automation. And that's how we have, that's where the industry is. So actually right now, even most of it. The third generation is when the automation is, you don't write scripts to automate things. You tell our system what you want to achieve and it generates automation for you. And that's what we call intelligent automation where it's all declarative and all the, you don't have to maintain a lot of scripts, et cetera, because they are, you can't keep up with it. You have to change the process all the time. And if you change the process, it doesn't work. It becomes completely, it becomes very fragile to manage it. So that's really where intelligent automation comes in. I look at like, if you can have like, if you look at like a Tesla, making cars, the entire assembly line is automated. But if you want to change something in the assembly line, even that process is automated and it's very simple. So it's, and that's what gives them so much, you know, let's say control and manageability around the manufacturing process. So the software delivery, you know, by assembly line, which is the software by CI CD pipeline, really, should be more sophisticated and more intelligent as well. Now then that's an acceleration. Yeah, Jyoti, you're also pointing out something that we cover a lot on theCUBE and we've been writing about is how modern software practices are changing whether it's team makeup or whatever. It's speed is key, but also getting data. Everyone who's successful with cloud and cloud scale. And now you got the edge opening up. And like I said, even space is going to be programmable. Everything's programmable. And the key is to get the data from the use cases, right? Get something deployed, look at it, get some data, and then double down and make it better. That's a modern approach, not build it and then rebuild it and re-tear it down and rebuild it. Which you're kind of leaning into this idea of, let's get some delivery going, let's structure it, and then feed it more so that the developers can iterate with the pipeline. And this, again, can scale. Can you talk about that? Can you comment on your reaction to that? Yeah, definitely. That's exactly how we look at it. Like, you want developers to kind of like say they want to do an automated process to deploy in their Kubernetes infrastructure. In matter of minutes, you should be able to get started. But now it's like, there's so much data that comes into it, like that you have monitoring systems, systems like AppDynamics and New Relic and Datadog and your logging systems, your Splunk and Elastic and Sumo Logic, you have your different kind of testing systems, you have your security scanning. So there's so much data in it. They're like, terabytes and terabytes of data from it. So when you start doing your deployments, we could also consume all of the data and see what was the impact of those deployments or code changes in each of these monitoring, testing, logging data systems and how the data changes. And then now it's like, based on that, we can learn what should be your ideal process and what will break in your process. And that's how Harness platform works. That's the core of that intelligent automation network. We are expanding it now to bring a few more of the DevOps use cases into it also. Like the one is cloud cost management. Because when we started shipping this, a lot of people would tell us like, you're doing a great job helping us managing the quality which we always were concerned about like when we're deploying things or security functionality, et cetera. But cloud cost is a big challenge as well. You have your paying like tens and tens of millions of dollars to the cloud providers. And when developers do things in automated way, it could increase the cloud cost suddenly and we don't know what to manage that. So that's the, we introduced a new module called cloud cost management to, as part of the DevOps software delivery process, that every time you're shipping code, can we also figure out like, what were the impact on your cloud cost? Can we automate the, if there is too much impact, can we automate the rollback around it? Can we stop the delivery process at that point? Can we help you troubleshoot and reduce the cost down? So that's cost becomes another dimension to it. Then we recently just added the next level that's managing feature flags. And a lot of the time software developers are adding feature flags to, like this feature would be given to this consumer and like, you know, and this feature would be given to this consumer until you test it out through a AB test kind of thing. And like, you know, what is the impact of, you know, turning a feature on versus off, you know, we are bringing that into the same CI CD pipeline. So it's kind of an integrated approach to this, you know, our intelligently automated pipeline instead of these small point approaches that just very hard to manage. I mean, the level of data involved, the feature flag, for instance, is great, is an amazing thing because that allows you to do things that we used to be extremely difficult to provision. I mean, just picking the color of icon, for instance. This kind of blue, I mean, I was just, you hear about these kinds of things happening at scale and the data's pretty accurate when it comes in. So I think that's an example of the kind of speed and agility that developers want. And the question I want to ask you though on that point, because this opens up the whole next conversation, you guys have a modern approach and so much traction and you've recently raised big rounds of funding. As you go to the marketplace, you're an experienced entrepreneur and CEO, you've seen the waves before. What's the big wave that you're on now? What's the big momentum tailwind for harness? Is it the fact that you're creating this value for developers or is it the system that you're integrating into with the intelligence to make things smarter and more scalable? What's the, or is it all of the above? Could you just share what that story is? Yeah, I think it's really both of them. Like, our business case, when you go to people, we tell them, like say, if you're 200 developers, we can give you the world's best software delivery tooling at the cost of half to one developer, right? So, which is like for a 200% organization at $200, $300,000 a year, they will get the best software delivery tooling better than a Google, Facebook, Amazon kind of companies. It varied very quickly, right? So, our entire value prop is built on that. Like, you know, your developer experience gets much better, their productivity gets much better. Developers on an average are spending like 20 to 30% of their time on deployment delivery related toile, like unnecessary stuff that you have to deal with. So, it's 20, 30% more efficiency gain for your developers. Like, their quality of life gets better, that they don't need to worry about like weekends and nights to babysit your deployments and, you know, things breaking and troubleshooting things all the time, right? So, that's a big value. But as a business, you get much more velocity. Your innovation velocity is much higher, you know, your risk on your, you know to your consumers is much lower because your quality of the, of, you know how your ship becomes better. So, our business case of like, you know at the cost of like one to two DevOps engineers will get you the best DevOps, you know tooling in the world possible, you know it's not a hard business case for us to make, right? So, that's what we look at and it becomes pretty obvious for, you know as people try our product, you know, the business case. You don't have to really pass the IQ test to figure this one out. Okay, everyone's happier and you have more options to scale and make more money in new opportunities not just existing business. I mean, the feature flag and these new features you can build new value and take more territory if you're a business or whatever your objective is. So, clear value. Can you give an example of some recent successes that you've had or traction points that you think's worth notable that people couldn't get their arms around? Yeah, definitely. Like, you know, we are helping a lot of, you know a lot of customers, you know doing like completely changing their process of software delivery. You know, one recent example is nationwide insurance. You know, nationwide insurance, you know moving from their data center kind of approach to public cloud and to communities and to microservices like a major cloud native rearchitecture and in a very ambitious aggressive project to do it, you know, in a short period of time and harness becomes a platform for them to kind of, you know to remove all the bottleneck around the process the software delivery process, you know they obviously they still have to do the developer side of things and they have to do the cloud infrastructure side of things which they're doing but the entire process of how you bring together you know, harness becomes the accelerator around it. So there are a lot of these kind of stories that we when we kind of create this fundamental transformation for our customers, you know you know, moving to a public cloud, you know moving to microservices, moving to Kubernetes, you know rearchitecting that they become much faster cloud native, higher, you know a true software company. And you know, I would say that's something we take a lot of pride in. I think our biggest challenge is to evangelize and convince the market that this is possible to do with the product, you know because historically people have been told like, you know, the only way you can do these kinds of software delivery processes and tooling is by engineering it on your own. So everyone was on the path of writing their own, you know and it's very hard for every company in the world to become very good in writing your own software delivery, tooling and processes and systems, et cetera So it's, and that's, you know there is still that education and evangelism needs to be done that, you know, there is no point you're trying to do it on your own you can get a platform that can do it all for you and you can focus on your core business of what you want to innovate on. And I think the DevOps movement has been pioneered and you had to hand roll everything and that's the way it was but now as the mainstream market picks this up you're standing on the shoulders of those pioneers you are one of them it's awesome to see this modern approach because it's really playing out in real time. Again, you've done that before Joe T. So it's impressive and you know you know you've seen the movie in DevOps and the earlier versions pre DevOps so as cloud native comes and starts scaling it's going to be for the rest of us. So great, great that you're providing the platform and the tools and the software. I got to ask you that if you don't mind because a lot of people are looking at ways for modern approaches to organizing their teams how would you define the modern DevOps movement? You look at DevOps 1.0 we got here, okay cloud, cloud native, cloud scale, modern applications, CISD pipelining now we're looking at a whole nother level of confluence of integration and speed. How would you define the modern DevOps movement? Yeah, I think that's a very good question. I think that the core of modern DevOps what I would call it DevOps 2.0 to me is developer self-service. It was like the first generation of DevOps was they create this kind of a DevOps team and then the developers will give all the delivery related stuff to the DevOps team and the DevOps team starts to become a bottleneck everywhere now like in the DevOps team job is to build the CI pipeline and the CD pipeline and the deployment scripts and do like you want to do a canary deployment they have to figure it out how to do it they have to do like you know you're in all sort of things that needs to be done you create a central DevOps team and you give it to them and they become a big bottleneck. We look at the modern DevOps or the next generation of DevOps has to be done around focusing on the developer experience and making it all self-service for the developers. So you have, let's say you are a dev team for a microservice and it's like five, seven engineers modeling a microservice. You want like that they can go and say this is for our microservice, you know in a matter of minutes or hours they can engineer the process without having to lean on a central DevOps team and to do all the work for them. And that's you know by maybe in a modular or in some kind of a YAML interface or something that's very easy for them their experience is so easy that they can manage it themselves without the central DevOps team have to write it all or code it all and manage it all. But at the same time the central DevOps team's job becomes of our own governance that can they define the guardrails that they can define the guardrails on like you have to have this level of security before something goes into production you have to have this level of quality before something goes into production you have to have like you know this your cost could not be more than this, right? So you define so instead of the DevOps team is doing all the work themselves on writing all the stuff they define the guardrails and it becomes a very easy self-service experience the developers should do things within those guardrails is what the modern DevOps should be. That's awesome. And also accelerate more business value and you're nailing it. Jyoti, thank you for coming on and great to see you on theCUBE CEO and co-founder of Harness. Harness.io you guys got free trials free downloads you got a great buy as you go model. Also you're an entrepreneur at heart co-founder of unusual ventures big labs, app dynamics now Harness congratulations. Thanks for coming on. Hey, thank you, John. Okay, this theCUBE conversation I'm John Furrier in Palo Alto, California with theCUBE, thanks for watching.