16 January 2024

Cervin Founder Spotlight: Ellen Rubin of Causely, Part 1

Ellen Rubin, Co-founder & CEO of Causely, sat down with Cervin partner Daniel Karp to share her experience as a third time founder and offer advice for the next generation of entrepreneurs.

Causely is bridging observability with automated orchestration for self-managed, resilient applications at scale. Here is part one of their two part interview. 


Daniel Karp, Partner, Cervin: Thanks for joining me today, Ellen! Tell us a little bit more about yourself, Causely, and your entrepreneurial path.


Ellen Rubin, Co-founder & CEO, Causely: I’m what you would call a very seasoned entrepreneur. Causely is the third company that I've started. I've always been intrigued with enterprise infrastructure problems. It's really interesting to see how things run in the real world and understand the underlying technologies that make that possible. That’s been a theme for me for a long time. I really like working with early adopter enterprise customers that are thinking about how they're going to build things for the future. 


A Founder's Origin Story

When I started off in my career, I was part of a company called Netezza, in the data warehousing space. I was among the first few people on the business side joining when we were still building a product, and that was pretty formative for me. At Netezza, we built a very advanced system that enabled people to do data analytics at 100 times the performance at half the cost. The customers loved the product. I created the go-to-market engine, we created a category, Data Warehouse Appliances, the company went public and ultimately, was bought by IBM.  


After that, I pivoted into the cloud, which is a career passion for me. In 2008 I started my first company, CloudSwitch, and it was all about hybrid cloud. Then in 2014 I started ClearSky Data, which was focused on storage and how storage can be used at the edge at much higher performance, all while taking advantage of the economics and the scale of the cloud. CloudSwitch was acquired by Verizon, and ClearSky Data was acquired by Amazon Web Services (AWS). So then I did a tour of duty at Amazon for a year and a half. 


DK: So you were acquired by Amazon around 10 years into the formation of cloud and roughly nine years after you started your involvement in the cloud, since AWS started in 2006. 


ER: Yes, and I started CloudSwitch in 2008. So at that time, there was no other cloud and Amazon was this emerging superpower. I've always interacted with Amazon as a competitor, a partner, a scary behemoth; and then I became part of it. I joke with some friends of mine that we all end up working at cloud computing companies. We’re either acquired by them or we become them. That's the era that we're in. 


And now coming to Causely. When Shmuel Kliger (Causely Co-Founder) and I started working together almost two years ago, the whole frame of reference that we both had was that the world is in the cloud. And that doesn't mean there’s not still hybrid with some things staying on prem, and that you don't still have a lot of legacy, but we're focused on the stuff in the cloud that is already mature, advanced, decentralized, and chaotic. Even though cloud apps are designed and built in order to get rid of a lot of the previous generations of messy, horrible IT operations, we’ve managed to create a whole new set of complexity and challenges in the cloud. So there's lots of room for a new start.


DK: Both you and Shmuel, you're part of that transition into the cloud, you've seen multiple markets expand and get centralized again in mainframe and then cloud. Causely starts as cloud based, but it also would need to take care of very complex workloads that are hybrid environments. And also, if you think about building applications in the cloud, those are kind of compartmentalized with multiple building blocks. So how do you think about that complexity and what inspires you to solve that huge problem that no one was able to so far?


ER: Causely’s vision comes from Shmuel and his background. He was the founder of Turbonomic and before that, he was founder of a company called SMARTS. He’s been looking at automating IT operational challenges for a long time, starting way back when we were still dealing with hardware on prem and then moving from there into virtualization and then into the cloud. We've both been on these journeys, over many, many years. 


The philosophy that he and I share is that in spite of how much automation and ability for people to see everything clicked on, and how prevalent observability tools are, which show everything happening – if you are running cloud applications, these lend themselves to being incredibly multi-layered in terms of the complexity that exists. And so we're starting off thinking about it in the context of applications being built using microservices, decoupled, highly elastic and very scaled environments. And so in that environment, the thing that turns out to be really hard is understanding the interdependencies between all of the disparate components. You could potentially have millions of interdependencies between all of the different services across all possible layers of the application and down into the underlying infrastructure. People who manage and run these applications, frankly, often only understand the piece that they're working on, or maybe they developed a particular set of services and they understand that. Some people are platform people that are closer to the infrastructure, other people don’t want to be bothered to talk about load balancers. So you have these teams that are troubleshooting and dealing with a lot of the complexity manually. What Causely is doing that nobody else does, is capturing causal relationships and causality across the entire application environment in our software. We feel that even though you have observability, and tracing and metrics and logs, a lot of what happens today is that people apply human knowledge and human effort based on their expertise to understand what's going on. What problems are emerging, are they getting into a situation where there could be failures, performance, latency, availability (issues), and they don't have a way to capture that into software which could then lead to automation, through prevention all the way to remediation. Up til now, knowing causal relationships requires human beings to do their best job based on expertise they have. 


DK: I think that’s because of the scale of the environment. I believe that the company labels the relationships as manifestations and attributes, while others call them permutations, but those are impossible to capture by a human. 


ER: Think about it almost like here in New York City. You're standing on the street corner and can see that there is something causing a delay or that people are not able to move, and you can understand what's going on around you. But you might not know that all the way down at the other end of the island, other things are happening and that’s what's impacting you. All you're seeing are symptoms.

DK: And that's what the industry is focused on - Just make sure that you're not underwater and the service is up and running, and you just look for the most probable cause, rather than trying to figure out how to solve the problem in its entirety. 

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Maecenas a fringilla tortor, et porttitor tort. Vestibulum non nisi interdum, blandit dolor in. laoreet magna. Suspendisse sit amet elit sit amet nisl. semper imperdiet. Suspendisse

The Emergence of Causal AI

ER: Right. A lot of the tools out there rely on machine learning and different types of AI, but a lot of it is based on correlation - things that are happening at the same time, or nearby and therefore may appear to be related and then they’re toggling between different views and tools. Even if they use tracing to get more context, that’s still limited in what you can do. You can show things are connected but you cannot explain the causal relationship. So what we are focused on this emerging area of causal AI, and capturing that in an automated way across the core platform that we've built, and then applying it to a whole series of problems. And a lot of the problems that we're starting with have to do with congestion, noisy neighbors, database issues, messaging brokers, you know, things can be slower and unexpected and unpredictable. Things that look like maybe there is a problem, but it's intermittent, and it comes back five minutes later, or it's an emerging problem. And so we're really trying to think about how to maintain the overall application environment in a good, healthy state. And if things start to go wrong, we're able to remediate them and tell you exactly what the root causes are and fix them right away.

DK: Shifting gears - many entrepreneurs in our portfolio are first-timers, but the Causely story is completely different, you are a third time CEO and entrepreneur, part of canonical companies like Netezza, and Shmuel is a seasoned entrepreneur. And moreover, five or six of the founding engineers were in Turbonomic with Shmuel.

ER: Some of them have been working for him for at least 10 years, and some of them 20 years.

DK: And so tell us a little bit about how this time around is different from your first venture and your second venture. What have you learned? 

Taking the Leap of Faith

In the first venture you have no idea what you're doing, or what you've gotten yourself into. I describe it as a leap of faith. You jump into the chasm and you're trying to figure out what's going on. When you've done this several times, like Shmuel and I, you want to be careful to retain the spark of excitement and serendipity that exists in that first time, all while leveraging the experience of identifying problems as they rise and the bottlenecks which could ensue. That said, if you thought about all the things that could go wrong, or the ways in which things are going to be challenging next year, you wouldn't do this, right? So you do have to maintain that leap of faith that you're making. We're trying to make sure that the people we brought to the team have that spark in them, and it's very helpful, of course, that the people on our team have worked together closely at previous companies, since we are remote. We also try to be really thoughtful about the way in which we are testing and iterating, versus in earlier companies, where we would build a product for a year or two, and then you go out and try it and see. You want to have that iteration mindset, but also that there’s enough clarity to bring it to market, and we're planning to bring product into the market in 2024. We all care a lot about the value that we're going to bring to the customers and that we are adding value right away. 

DK: My observation is that you are very focused on the customers, not just building the product for the sake of solving an engineering problem. The second observation is that there’s less drama - it feels like things at Causely are balanced, which is great, because there's the nuanced understanding that there are a lot of things still being figured out, but that’s ok at this stage.

ER: Shmuel often encourages us to take a step back and sometimes be okay with a lack of clarity on things because we're on a journey and we're figuring it out. We’re creating something that's truly breakthrough technology. No one has applied causal AI to IT operations - that has just not been done. We're thinking about how to work with customers to build a product that provides them with real value. We’re learning things through our design partners that are really helpful to us. We’re also thinking about how to grow the team, as there’s more clarity about what to build and to do, we think about what skillsets we still need to bring to the table and how we build a team in a remote way all while the culture is strong.

Stay tuned next week for part two of this interview!