Unlocking the Potential of AI in DevOps – Kumar Chivukula, Opsera

This is Textron tv. Hi everyone. Welcome back here to Techstrong tv.

You know, I'm really happy to have our next guest on. They, he came on my show, he came on our show here at Techstrong tv. It's gotta be three years ago, maybe new company that just kind of come outta stealth and everything.

And we were really happy to have him here. Um, let me introduce you to Kumar Vu Trala. Trala trala co-founder c e o of Obser.

Kumar welcome. Pronounce your name correctly for me. I did a terrible job.

You've done well, thank you. And you, uh, really appreciate, uh, to have, have, have us here, have me here, and, uh, uh, honored to be, uh, talking to you. Glad to be here as well.

And we met a couple of years back when we raised series around, and then, uh, we had a great conversation and, uh, happy to join you back again. Talk about the, uh, recent developments that happened in the in ops Ops. So, You know what, Sera made an impression on us.

If I'm not mistaken, you guys had won a DevOps dozen award when you, you first came out as a new company, and yes, it was good stuff. Um, but you know what, not everyone in the audience is really gonna know the Sera story. Maybe they haven't heard of Sera.

So, Kumar, if you wouldn't mind, give people a little bit, maybe a little bit of your background, your history, and then really let's talk about what, you know, what if Sarah does, and, and it's kind of the Sera story, and then we'll jump into today's news. Absolutely. I think that that's a great way to start the conversation.

So, uh, Sera, we are a unified DevOps platform and, uh, we, we think about it the way it used to be called, the category that was created Gartner recently and previously we used to call as a value stream delivery value from delivery platform. And we've been working with, uh, industry and the team for like, for almost two and a half years. And we started as a DevOps as a platform.

We are so glad that, uh, the Gartner changed the category to DevOps as a platform. We've been included as honorable mention. So that's a good way to start this year.

And then, uh, glad for our mission and vision has not changed. Our goal and mission is continue to stay true for us, and, uh, we want to help enable empower enterprise companies to make the software delivery faster, better, secure. That could be any software, could be soft, S D L C, which is Software Development Lifecycle, which you see the traditional workloads with the various languages and SaaS applications, which are Salesforce, Informatica, snowflake, Adobe Experience Manager, and so on and so forth.

Because a lot of the IT companies have spent a lot of money on SaaS applications. They gave up the internal development to SaaS, but now there's so much development happens in SaaS applications as well. The third one is infra as a code, which are the three main use cases, combination of that where we are bringing it together, all three use cases with unified insights, where in which we can provide that governance and visible end-to-end visibility for the enterprise customer to make smart decisions.

That's a quick summary background. We started in 2020, came out of stealth in 2020, uh, October, end of October, November, and, uh, for the last, uh, two and a half years. And, uh, we've been working with a lot of, uh, fortune 500, fortune hundred companies to transform their DevOps and DevSecOps journey.

Absolutely. Y you know, we, we, uh, just yesterday we had our, well, yesterday from when we taped this last week from when people see it, we, um, we had our DevOps, enterprise, DevOps, not DevOps, enterprise Summit, DevOps experience, virtual Summit. And really the theme was achieving balance in DevOps.

com 10 years, uh, when I first started, I, I look back on it now and I realized maybe I was naive in terms of how simple I thought of what DevOps is and, and how to implement it and, and you know, what it can do. And of course, it's, it's grown up into this over the last 10 years. There's so many different aspects to a successful DevOps and DevSecOps, which, you know, that was one thing I did have, right?

I knew dev sec, I knew security was gonna be a big part of it. Yeah, Absolutely. Group prediction.

Yep. And, and that, you know, and I look at DevOps DevSecOps now and, and, and things like platform engineering and ss r e and observability, and of course generative AI and low code, no code. There's all of these macro kind of things that are pushing and pulling and stretching that for many organizations, especially larger organizations, achieving the right balance, the right mix in their DevOps strategy is, is not easy, right?

It's harder than ever. Absolutely. Because you, you're not the only one.

I think you've seen the data reports from the both Gartner Forester and other people like Gig Home and whatnot, that that's one, one aspect of it. We also, I can also share some of the experience, some of the nuggets that we, uh, we see from the, our interactions and conversations. So 87% of investments and, uh, the, the enterprises made in the, uh, DevOp space did not pan out because it's all people are approaching the point solutions and, uh, now the people, people are adopt as a platform and try to expand with keeping the autonomous of the developers and giving the control, giving the freedom, and having the feedback mechanism with automation, visibility, governance.

That's where you see the success. And, uh, they're also predicting that by 2027, right, over 20% of people adopting the DevOps as a platform, and now 2027, the Gartner is predicting to have about 70% of enterprises will adopt DevOps as a platform. So those are, and That, that's critical mass, right?

I mean, 70%, anything over two thirds like that is, I, it's a critical mass number. Fantastic. So, and look ER's right in the middle of this, you guys recently announcing some news here.

I'm not, I don't wanna steal your thunder. You tell 'em what, what's the news? It's a, it's a collective team effort.

And, uh, because we have to have a great team to get to where we are, and, uh, we are fortunate enough to have, uh, wonderful teams supporting us along with, uh, uh, customers like, uh, the Fortune a hundred, fortune final customers, and where they've been, uh, been part of our journey and we thank them. And also investors, right? So as part of our, uh, uh, success and also like what we've been, been, what we built as part of the, our Sera, we are able to raise about $12 million and series c plus town.

And, uh, we, uh, lead invest was Taiwan in capital and followed by the, along with, uh, our clear ventures, we, uh, we were able to finish the close around. And our goal is to invest the money more towards accelerating our growth sales growth and go to market side. And the further we want to invest in three prongs, and one is, uh, put the money into the go-to market where we see the significant, uh, demand there.

The second one is around product let growth. How do we make it more easy for people to use using the product? Let go p l g motion and which is what the natural evolution for us, where we start versus where we are.

Third one is our generative way on r and d effort, right? So generative way, the conversation does not start and end nowadays without talking about generated ai. If you bet on it, and you, everybody makes money on that.

So, so our goal is to talk, take the different approach in generated way, which we can talk about it and uh, uh, in detail. Our goal is to take a, since we approach as a platform, we want to apply the generative AI initiatives, uh, into the platform as opposed to like just randomly add some things to it. And we are taking three prong approach for that.

The first one is we want to bring the, we have a lot of patterns and uh, we, we've been, we filed some time back and we got a bunch of them in the predictive insights, anomaly detection, pattern analysis. So you wanna make it, make use of them. So as a result, we designed our, uh, the overall platform and then generative initiatives into problem-based engineering is the first one where in which you can get your insights of developer productivity, developer experience, security metrics, quality metrics across the board.

And just a matter of just fingertips, you just type a question, how many pull requests on this particular release, how many deployments were done, how many, uh, uh, your bugs are there? How many security issues? It's just the simple questions, but it takes so much time to get the data as a C T O, imagine that you're a C T O.

You want to know where, where the velocity, agility, change, failure rate, deployment, frequency, uh, your, uh, your lead time with the DOA metrics. You don't have to worry about pulling your hair. You just come to Sera, connect to your sources and type the questions, you'll get the answers.

That's the first way, problem-based engineering. Second part of the conversation thing, what we are doing is how can we, em empower developers and DevOps teams to do things faster, better, and secure? So as a result, we integrated with the, uh, or some of the large language models and, uh, we created our own, um, model ourself within the framework, I should say with that will help and, uh, the developers and DevOps teams to troubleshoot issues faster without them worrying about, okay, when there's a 4 0 4 error, when there's a, uh, page, when when the docker image is not there, or maybe some deployment failure or bill failure, may one or greater, we will give you the remediation resolution.

Only thing we don't give is we don't allow them, we don't give the button, uh, to them. We don't do the self self feeling right now, but we will eventually, we are given the button for them to press it by just looking at the information, right? So that's the second part.

Third party is about like, it's a co-pilot as assistant, whatever we want. We, we are still finalizing that in terms of like where you can, uh, create a, using the platform. Imagine that you're a developer, you write a, wrote a code in Java, go language.

We just want to take the metadata and be able to create the pipeline in the background without worrying about which pipeline to use. Where do need to deploy? All I have to say that I need to deploy this, my Java or go language container and do a w s Azure G C P, we'll take care of the rest in the background and, uh, we'll give the result and the encapsulating all of them.

We are also trying, enabling the enterprises to deploy the large language models. And that's another thing, new dimension that we are adding to the platform. Where in which not only do we deploy the model, you can understand the cost basis.

We can understand the efficacy, you can understand the bias scoring. So that's the additional value we want to bring in. So combination of problem-based engineering l l m model deployment, along with, uh, enabling developers to do things faster with respect to pipelines, and then having the co-pilot assistant that the combination of that we calling a Hummingbird AI platform.

Hummingbird ai. Love it. I love it.

That's our logo. People are familiar with the Obera logo. It has that hummingbird in there and it, it just makes a ton of sense.

You know, Kumar, you, there was a lot of ground you covered here. I wanna, if we can break, break things down. First of all, congratulations on the money raise in today's market dynamics.

Raising money at all is hard. $12 million is a very healthy round. And congratulations to you all.

Thank you. Uh, in in raising that. Yep.

Um, in, in terms of, of what we see in the, you know, 70% you mentioned is, is a number in maybe two more years, three more years, well, four years, 20, 27 of, of market adoption for DevOps and AI is, is such, is such a wild card in this that we're starting to see. I mean, some of the things you're mentioned custom, L l m, right? I think of that as the short term memory that gets put in underneath the underlying, you know, maybe larger dataset, L l M, but it pulls some short term first.

I think these, those are the kinds of real world examples, real world solutions where it kind of, you know, don't believe the hype, yeah, it, you know, about ai, but these are things we could do now that really do help that really move the needle that really make a difference. And, and I think it's, it's easy to talk about what could be, will be, should be and, and, you know, writing code by itself and everything else, but where's, where's it, where's it real now? And, and I think the, this is a great example.

So is this Hummingbird ai, is this now like part of the core op Sarah offering? Is it a, a separate offer offering? What, what do you see there?

Great question. So it's, it'll still be part of the platform, but it's an add-on that people, uh, want to use the take advantage because it's, it costs money and the resources and it's not easy to sure deploy and implement that because we learned a lot. We've been working on this one for quite a, quite a few months.

We, uh, like I said, we have filed the patterns. The only thing that changed in our world is the models that you talked about it. I didn't have to, I don't have to create a model myself, but you still have to train the model.

You still have to validate the model. You still have to deploy the model within our own ecosystem. That is not changed in the, in the current landscape.

But what has changed for us is like taking advantage of the data sets that already created and, uh, how do we train them? So to me it's the, the cost of running the models and the, the you, every time you make a query, it's a token now, and it's all about token based. It's no longer ca utility based computing, it's token based computing.

So mm-hmm. Tokens cost a lot of money behind the scenes and without you knowing it, it, uh, catches up on the new bill. So we want to see how do we, we learned a lot about the models.

We looked at multiple open source models and whatnot. So we landed on the stack that we think it's going to work for us in the next few years. And, uh, to maintain that, manage that, it's an add-on con add on thing for the enterprises, but also it unlock the potential like enterprises by using ai, they can unlock the potential where then they don't have to dabble around and, uh, they can protect the existing investment using Opsa and then be able to use Hummingbird AI to unlock the potential that they're not able to do it.

Fair enough. Um, you know, I, I wrote an article last week around something amiss or changing in the open source world. We're seeing companies change their business licenses, you know, the software licenses away.

Yep. From open source, we're seeing companies pull back their financial support of some of the larger, like, you know, foundations and open source kind of managing management. We're seeing companies move away from offering open source, uh, parody with some of their commercial products.

I'm wondering in in, in the case of Sera Hare, how does this play out? You know, I mean, as you said, nothing's for free. All these things cost money, tokens cost money, doing ai, you know, and the re and the, just the hard engineering involved in bringing a solution like this out.

Do you think, Kumar, do you see a, a wind of change maybe in, in what, you know, people's expectations around what's free and what, what's value and, and what they should pay for it? Uh, it's a great question and, uh, I, I will have to take Op hat off for a minute and then talk about that. Uh, because it's a, it's like lot, lot happening and that open source communities, in my opinion, personal opinion, open source community is gonna thrive even more than ever before.

And, uh, especially mm-hmm. Uh, the models and, uh, the being created in ai, it's going to like, give more opportunities for the developers to create things that we've never able to create before. And, uh, you see more developers coming into ecosystem as a result, you get a different set of open source technologies that wanna pop up and, uh, which you are struggling to create before, like, like I remember the days where we had to compile entire Linux and there things that we had to do with Fedora and Red Hat and a bunch of other things.

Obviously Red Hat showed the path, uh, for people and then, uh, various, various other open source communities followed through it. I still believe that open source is here to stay in investments will be there, and it's a new wave of investments. I think what we have seen in the, it's maybe like what we saw 10, 15 years back.

The community, open source community and open source technologies are going to transform the, the technology is going to change. And we are here to see the transformation of what's the new thing that is coming in. But I firmly believe that open source is here to stay.

And even though one or two companies are changing the investments for the, but they're not changing the model though. What they're changing is when people are taking advantage of open source and try to create the companies, but competing against the business for them, that's again, the way I understood they want just make sure that they get the fair share. Either you contribute back to the community or you just pay the license fee.

So we are not here to violate any of them. We, we honor that, we respect that we continue to draw the same fine, keep the fine lane, make sure we continue to contribute and contribute to work, to work on the open source community. I love it.

Hey, you know what? We haven't told people. Where can they go get information about Hummingbird ai?

Yeah, go to t io www dot t io slash hummingbird and you will find that in the top of the banner page. And, uh, we have, uh, some exciting offerings and if you are ready to try it out, please uh, sign up and be happy to work with you and see how we can enable the Hummingbird AI figure. io and it's slash Hummingbird if, if you're interested in Hummingbird.

But you probably wanna start with the whole Obser platform and get a flavor of that as well. Kumar, I want to thank you for coming here on Text Drunk TV with us today. Don't be a stranger.

Keep us posted. Best of luck with, with Hummingbird ai, as well as Sera General, and maybe we'll see you, I don't know, maybe at reinvent or something. Absolutely.

Thank you for having us and having me. I really appreciate it. Thank you.

Your time and insights. My pleasure. Have a good All.

Bye. Absolutely. Kumar Akala, co-founder, c e o of Sarah on their new Hummingbird io, as well as a $12 million raise here on Tech Drunk tv.

We're gonna take a break. We'll be right back.