Podcast #144: Edge-Native

This is Ken Forster, Executive Director at Momenta. Welcome to our Digital Thread Podcast, produced by, for, and about digital industry leaders. In this series of conversations, we capture insights from the best and brightest minds in digital industry, who are executives, entrepreneurs, advisers, and other thought leaders. What they have in common is like our team at Momenta they are deep industry operators. We hope you find these podcasts informative, and as always, we welcome your comments and suggestions.

0:32

Good day, and welcome to episode 144 of our Momenta Digital Thread Podcast Series. Today is my great pleasure to host Rachel Taylor, CEO of Nubix, an Edge-Native application platform that makes it easy to build, deploy and manage applications, even on the tiniest of edge devices for the IIoT. Momenta is a proud investor in Nubix.

Rachel is a passionate operational executive with more than 20 years of enterprise infrastructure technology experience. She has held numerous leadership roles at leading tech companies where she's helped build great teams to achieve operational excellence, as well as various liquidity events, including Cloudera, Meraki, PeakStream, Riverbed, and VMware. Prior to Nubix, Rachel was Chief Operating Officer at Rocana, focused on securing funding and scaling the organization.

Rachel, welcome to our Digital Thread Podcast.

1:27

Thanks, Ken, it's great to be here.

1:29

And it's great to have you. I always end up saying the same thing at the beginning of this, we should have done this a long time ago, and it certainly applies to you because we've been an investor for a while. I now have worked with you in several different capacities, and I know how hard it is to get your time, so I really greatly do appreciate you taking the time to talk with us today.

So, as I always I like to start this to understand your own digital industry leadership journey. What would you consider to be your digital thread, in other words, the one or more thematic threads that define that journey?

2:06

One of my favorite sayings of late the last couple of years has become that life only makes sense in reverse, so maybe that's because now I'm in my 40s, and life is starting to make more sense. But for me starting off in tech, especially 20-some years ago as a woman, was really challenging, and there were a lot of roadblocks, but I navigated. I continue to tack, to use the sailboat analogies. I just kept tacking and tacking, finding the wind and current. Eventually, I made my way back, but it's been a strange journey that now I absolutely am so grateful for the turns that life gave me along with my professional career that landed me here.

That really has been starting off in technology, where I wrote my first program in second grade and was just hooked on technology. I went to college early. At the age of 16 was the first sales rep on campus for Apple for their H2 program. I got to go to the Irvine campus of Apple at 16. I had to get a waiver signed in college from my parents. I worked in the computer labs all through college, then started my first job after graduating from college at 19 and ended up implementing the first Windows and T-network in the city of California for a brokerage firm where I really solidified my love of technology for its practical ability to solve really-really hard, really important problems, in meaningful ways.

Then my journey kind of took a turn as I moved from Santiago up to the Bay Area; I thought I'd get a tech job and be a network admin, but instead, I got forced into technical recruiting and helped build five or six NOC centers for Exodus Communications before they blazed out, and like you mentioned I went to some Marquee Technology Companies really early and got to help them build those companies with a great technology idea, some money from some VCs, we got to go build the team and execute.

So, I had this front-row seat over and over again to what that process looks like and seeing the good, the bad, and the ugly firsthand. Sat in a lot of boardrooms, presented to Vinod Khosla and Doug Leonie and a whole host of the VC elite in Silicon Valley, and learned a lot about how to build a great company and how to make a lot of great mistakes along the way, and how to navigate that as part of Nubix's journey.

4:59

It's interesting, we have a well-known exec search practice, and so one of the first things I keyed in is the people leadership roles you did, both recruiting key talent and certainly managing them. It was interesting to hear you say you were forced in that direction because I think what a lucky break early on to get that. Because I know how one how difficult it is to recruit people, but two how empathetic it makes you toward understanding people and truly getting the most out of them. I'm actually fascinated by the technology and the talent acquisition space. I wanted to ask what attracted you to that intersection. I think you said you were kind of forced into it! But what were some of your key lessons learned out of that during that time?

5:46

What I learned is a couple of things. First of all, I always had a really great nose for early technology. I love to talk about the fact that I went to Riverbed before there was a WAN optimization Magic Quadrant with Gartner. I went to VMware before there was a virtualization Magic Quadrant. PeakStream got acquired by Google to power their browser at 20 people. I was at Meraki when it was 19 people and helped grow it to over 100 and still is the most successful acquisition Cisco has made. So great nose for tech, and tech that's meaningful, but also Tier 1. What I learned is first and foremost, go to the leader, don't go to the Tier 2 company, go to the people that are absolutely top shelf all the way. So that's number one, be the best, don't be a follow-on, don't be a second player, and they call that the kings of category. So that I learned.

The second was that A players settle for nothing less, again that top-shelf kind of attitude, like B players, hire C players and build C player teams. With A players you get that multi-player affective 10x, and that is massive. The other thing about people that I learned is that you want missionaries, not mercenaries. Mercenaries will be gone the second something better comes along, $2,000 more, whatever it is they'll be gone. You hire the people that are passionate both about what you're trying to solve but also about what kind of culture and what kind of environment, what kind of leaders you're committed to having in that organization.

Then the last thing I learned about people that consistency is king. Just like kids, just like dogs, just like relationships of any kind, romantic or professional, you have to do the things you promise early on, or else you end up with either people that are disgruntled or apathetic and just checkout, and in fact may be worse, are actively negative influences in your organization. Or you lose them, and then you end up in this kind of leaky sieve situation. So, a great idea without the right people to execute is not going to reach the kind of success that the companies I've been at have reached.

Then I'd say the last piece that always does the right thing with people, it will always work outright, even if that means letting somebody go, even if that means not promoting someone because you don't think it's the right choice. You have to be consistent with your values because it impacts so many different pieces of your organization's potential and opportunity, bandwidth, and overhead, and how fast you can move. It has been the most important thing that I learned on my journey to becoming a CEO is the people side of it.

8:52

Yes absolutely, again I think you were lucky or blessed to have such an opportunity so early on -  to have a foot in both, and we'll talk in a minute about how you're applying that.

I noticed a very interesting, I'd say, a real step up on your CXO career, so you left I guess, Cloudera and went with some of the execs to form Rocana in 2014 and took on the role of COO at that point. Just looking at the fundamentals, it looks like an incredible journey. You did a funding round in 2015 led by Google Ventures. You exited toSpunk in 2017, a typical Silicon Valley kind of experience in such a short time. What did Rocana do, and what did you learn as part of its leadership team so early?

9:39

Well, Ken again, it's funny, not to be a broken record – but everything ends up making sense in the end. I look at the companies I've been at, and WAN optimization actually has threads. In fact, somebody said the other day how some of the stuff we're doing at Nubix has flavors of Riverbeds WAN optimization piece of what we're doing on these edge devices. Then you have VMware, which is virtualization, which is the first version of early Gen-1 of containers is virtual machines. Then you get to Cloudera, which is big data, which obviously is core to being able to collect data, build models and build ML, but then goes wherever it needs to go to be applied. Rocana was the last piece of that which was, what we were doing at Rocana was collecting all of the machine-generated data, and infrastructure and IT infrastructure – so everything, your firewall, your routers, your servers, even your cooling systems – anything that's inside of that cage or connected to your IT system in that distributed infrastructure. We were collecting the signals, and then we were applying analytics to understand whether there was a denial service attack going on and how do you automate that? This is why Splunk bought us because Splunk is all about THAT infrastructure management and how do I know when things are going wrong across my enterprise infrastructure.

So Rocana was applying machine learning to those machines generated signals, which now I'm at Nubix it's like okay – we've got the WAN optimization piece, containers which are like VEMs, we've got big data which is Cloudera, and the last piece of the puzzle is sensors which are the equivalent at the edge of a router or a firewall. It is generating a signal, it's not human-based, it is collecting this data on a regular interval, and how do I take that data, correlate it with other signals and try and understand in some automated ML Data science way what is happening, and what action do I need to take.

11:50

What a great story in terms of how all of this converges to your leadership journey at Nubix, but obviously your technology accomplishments there as well. You founded Nubix in 2017 with your now CTO, Mike Gray. You already talked a little bit about the problem you set out to solve, but tell us about some of your key use cases and wins.

12:14

It's kind of funny because that's a slight misnomer is that I actually am not the original CEO of Nubix. I am the clean-up crew, I came in, Nubix was basically on its last legs, and I was put in as CEO by the board after about two or three months. I came in as COO with the expectation I would succeed the old CEO, but it was a two-to-three-year track. Within two to three months, the board stepped in put me in as CEO, and it has changed in so many ways it's not even the same company anymore. My original investor calls me the Phoenix because, from the ashes, I have built Nubix. There's not a line of code left from what Nubix was before, but what I saw was the right ideas as to how do we bring intelligence in an agile, safe way to the edge of the industrial environment.

Mike, my CTO, and co-founder and I like to talk about our experience in enterprise infrastructure and how did the cloud and how did technology that was built to solve these problems in large distributed enterprise infrastructures… when I was a Windows NT network admin, I remember all the pain of printers that wouldn't communicate, and oh I can't get to my email, why is this not working, what version of software is running on which servers, and all of that overhead, and now you've got the cloud. I swipe my credit card, I'm up and running on NAWS EC2 instance in a couple of minutes. That is what needs to happen at the edge, and that's where Mike and I's background…

Mike is much more services, databases, appliances, and all of that on the tech side. So, it was this beautiful marriage of services and IT infrastructure and how do you bring those learnings to this edge environment, because we saw that was the next wave, is that industrial needed help starting to solve these problems. So, we brought those learnings from the cloud, but we focused on what was true about the edge, its small compute its disconnected, hazardous, or rugged environments that have low power and all of these constraints. So how do you solve those problems in an edge native way but using the same kind of concepts that makes sense from that journey for IT and Cloud?

15:08

Interestingly, I'm thinking of you like the stealth turnaround CEO now! The Phoenix, actually, that's a really good one to potentially subtitle this one as, which is great!

We've talked in the past about IIOT in some sense. We think it is a way to describe the virtualization of OT, much as the virtualization of IT has happened over the last couple of decades, and of course, going all the way down to control levels. It's interesting to see how you've applied that from the enterprise IT side to the industrial side now. The industrial side, of course, is very much about deterministic life safety, hi-resolution if you will control many times. So, you'll know we're an active investor in companies like yourself, Edge Impulse, mutable, all of which are providing edge if you will capabilities for various fashions.

When you think about edge computing relative to IIoT, how do you think it differs from enterprise IT as an example; how do you design differently, how do you take to market differently for the industrial?

16:24

I love that question because it's one of my key go-to phrases that I like to talk about, especially when I'm talking to investors or senior business execs, and that is that I think one of the reasons why this problem hasn't been solved… I mean, GE Predix spent billions of dollars. I have a very good friend that was a senior VPO of product there, and if GE Predix can solve it, what makes Nubix think that we, with our little band of merry men, are going to go and solve this.

I think one of the reasons why we've been successful in getting the traction we have and getting the attention we have, is that we looked at what was already true about the edge, we kind of accepted the reality of the edge, a lot of people are 'Hey, just get a Greengrass device and slap it on there, get a GPU, but you think about oil and gas rig that has 30,000 data collection points and guess what, about two-thirds of those already have a sensor connected to a microcontroller, there's compute there – it's not that there isn't compute, there are a trillion microcontrollers out in the world today that I just got off the phone an erm, they shipped another 4.4 billion microcontrollers last quarter, it is only accelerating.

So, there are microcontrollers out there, but the answer for a lot of people is, 'Well, we'll slap a GPU on it,' you can't put 30,000 GPUs on an oil and gas rig. That's not even tenable from the perspective of electricity consumption. So, we really accepted the physics of the edge, small devices, MCUs. We're talking 256K of RAM, tiny-tiny devices. Maybe they have connectivity to the cloud. Maybe they don’t, a lot of them don’t, so they’re disconnected environments. Maybe they don’t even have anything but a satellite that goes over once or twice a day. They are maybe on solar or battery, so these devices come on, read a couple of things, turn on a light and then turn back off. There were just a lot of things about the edge that are so different from the cloud where I expect to have always connected, unlimited compute, anything I want is at my fingertips, this lavish existence in the cloud of over-consumption, over-compute, anything I want it’s kind of gluttonous.

But way out the edge, it’s like sub-Saharan Africa where you’re eking out whatever you can on these tiny little devices, and so Nubix really accepted the edge based on what was already there so that we could address all of that brownfield opportunity. It’s going to take time, but we really focused on how do we leverage the compute and the connectivity, and the power and everything that already exists out there, to be able to address those? And then as greenfield stuff comes online and there’s more opportunity there, that’s also addressable, and it’s this bringing it all up together, and bringing it all online into one platform, instead of ignoring the trillion microcontrollers that are already out there.

That’s where we came up with this Edge-Native idea, you’ve got cloud-native technology that totally makes sense in an unlimited compute, unlimited power world, but out at edge where you’re in Edge-Native territory, you really need technology that was purpose-built for what exists at the edge.

20:11

Let’s drill down a little bit and go back, if you will, on the use case examples. Let’s take the oil rig example that you have. How would somebody actually utilize you in that? Where do you see the largest impact of what Nubix is able to bring into an example like that?

20:31

Well, let’s talk about heat sensors. So, let’s say for an oil and gas rig, out of those 30,000 data collection points, let’s say a couple of hundred of those devices are heat sensors. They’re managing all different pieces of equipment, and maybe some of those collections need to have one particular algorithm applied to the data signal. So, I’ve got a sensor, I’ve got maybe a microcontroller, most likely two-thirds of the time there’s going to be a microcontroller attached to that. Right now, that microcontroller is like old phones before smartphones came along, where I can turn a light on or off, maybe I can shut something down, or maybe I can turn on the fan, but it’s pretty limited in my ability to address the situation, make sense of the data that’s coming off of that sensor, in a meaningful way. I may turn the light on and say, ‘Hey, I’ve got an alert, something’s overheating,’ but I, as the operator, am now going to have to go on a journey to figure out why is this light on, what is overheating, why is it overheating.

If instead of using Nubix, once I’ve installed Nubix as my container on those heat sensors, now I can apply, I can deploy analytics that are very specific to understand this data signal maybe even correlate with not just the heat sensor, now I’m going to correlate the heat sensors signal with a vibration sensors signal, or maybe a humidity, or some kind of friction. Then I can start to make sense of what’s happening, and step one is maybe not to take action, maybe step one is to inform the actual operators that yes, the heat sensor has an issue, but I can tell you what the issue is now.

Then as people get comfortable because you’ve got to walk before you can run, as people get comfortable, the goal is to get to the point where now I trust that interpretation of the data and understanding what’s actually happening. Now I can take action and let the machine start to be autonomous. That’s all really hard to do because across these heat sensors. There are going to be all different kinds of OEMs, different communications protocols, different chipsets. Right now, what that looks like is every time I want to update one of those sensors, I’m doing everything in CNC++, its slow, I’ve got to design an application for each individual incantation of that set of parameters; what’s the RTOS what’s the chip, what’s the board, what’s the communications protocol, what’s my I2C bus setting.

Those are all the things that make that really hard, and so what you see is those firmware devices just don’t get updated, Schneider updates them once every six months, and ABB didn’t touch them in nine months because it’s just too hard, and then we haven’t even talked about the potential of breaking those devices. That’s going to become more and more front and center as you see things like the corneal pipeline cyber-attack, now that was on the financials, but the hackers see that these edge infrastructures are great targets, because this firmware, these MCUs are just very vulnerable based on current technology stacks, and that has to get addressed, across water treatment facilities, streetlights, and nuclear facilities and things. That’s the next infrastructure issue that is pending for our country and for the world, and it starts at these MCUs where the sensors are.

24:23

Truly industrial edge in that regard. We have an investment thesis called ‘Brown is the new green,’ i.e., a brownfield is where all the money is at. Show me a greenfield in the IIoT for the most part. They’re very rare. And even then, because it’s probably the same corporate owner, they’re going to replicate what they already have volume purchasing agreement on all the other plants with. So, it’s a way of perpetrating or perpetuating just the legacy. So, it’s an interesting model, and I see your value proposition very clearly in terms of how you guys help both homogenize and then harmonize all of that together.

I like the people topics, so let me go back for a moment; I think I was joking with you before the podcast that if fortune-tellers were accurate, they should be, by definition, the richest people in the world because they can fortune tell everything! And in some sense, Exec Search Professionals, recruiting professionals should be the best at identifying and recruiting or retaining really key talent. So, I’m curious, as a CEO and clearly, the term the Phoenix goes a long way here, but what has been your experience in that regard; do you feel that because of your background you’ve been a better people leader, recruiter, and retainer of talent?

25:45

Well, I think you’d have to talk to the people this is referred to them! I’m a good people leader, I try really hard, I’m very-very focused on culture and authenticity, and the biggest one is, do what you say and mean what you say; don’t say one thing and do something else, it makes everybody mad, and it never goes well. But a couple of lessons I learned along the way, one of the companies I was at was backed by Vinod Khosla his admonition to the CEO was, ‘No matter what, you spend 25 percent of your time as a CEO on people,’ whether that’s recruiting, managing, mentoring, whatever it is, and that really stuck with me that even at the CEO level back then, and this would have been 15 years ago, that it was that important from his perspective.

The other thing that really stuck with me and I saw over, and over, and over again done wrong, and you see it in books, that the number one regret most CEOs have is not dealing with the people issue when it’s time to get somebody out of the organization. I saw that first-hand. I saw a company completely implode 65 people because of one bad person's choice that just did not get addressed. So that’s something that my team, I think, would say, that the thing they see about me is that as much as I love people, as important as people are, I’m also absolutely fearless in the face of the hard choices, you have to be willing to make the hard choices, which sometimes sucks right at that moment, but a great hire is absolutely as important as knowing when it’s time to get somebody out of the organization.

So, it’s really that balance, and it takes constant attention to the signals across my people of what’s working, what’s not working, and that’s the only way you keep good talent. We’ve hired two amazing engineers, and I love the fact that we’re so efficient; Mike interviewed two people, put one in front of me, we hired him, and he’s been phenomenal. For the other position, I think he interviewed five people, put two in front of me, we hired one, and he’s been phenomenal. The amount of overhead to find awesome people but then actually get them to stick and be productive is so low here, and I have a great network to pull from because I have a great reputation of having made the right choices. Early on in my career as a recruiter, I realized even if I had a fee attached to making a hire, if it was the wrong hire, I would talk people out of it, on either side, the candidate or the company I’d say, ‘No, this is not the right person. Yeah, I’m going to walk away from a $10,000 fee, but it’s the right choice long-term.

I built my reputation on those kinds of choices. So, I think all of that comes through and makes the job of hiring and building an awesome team actually really easy. It’s not as hard as it seems – it is for a lot of other people, but it’s because I think I try and head off those challenges early on and then deal with them swiftly when they do rear their head.

25:45

I think as you referred to it before the podcast, you called it your superpower, which is apropos. Well, thinking of your other superpower, one we’re very proud to be a woman founder, especially in this deep tech space, but sadly women still remain very under-represented as founders of tech companies. To what do you attribute your success, and probably more importantly, what would you advise to other women following in your footsteps.

29:55

I’ll go back to the fact that this has not been a direct path, there is no lay-up here, there is no clearly defined Point A to Point B, there’s no straight line, but I think that I was tenacious, one of my mentors is a Stanford Professor in the Business School over there, who says ‘Try another play.’ And I think that mentality of being passionate, knowing what you want, not giving up, trying another play, constantly tacking, looking for that opening, and that opportunity is all that is required to be able to sit in the seat I’m in now.

One of my investors his comment is, ‘I wish that more of my CEOs had 10 percent of the gut instincts, but the willingness to trust and be passionate, focused, and convicted, but still willing to constantly take in new information and adjust; so not blindly committed. And that is really hard, and it’s hard to nail that and thread that needle because you have to believe in yourself but have enough humility to not have what I like to call happy ears, where I’ve already made up my mind this is the path, and no matter how many signals come in telling you that’s not going to work, you just continue blindly forward. It’s hard to nail that right.

I think for women, it is an amazing set of skills if you’re willing to stick with it, and not get discouraged, and continue to believe in yourself. They say that being a CEO is the loneliest job, I like to say being a female CEO, especially in tech, is definitely the loneliest job. But you’ve got to be able to stand in that and be comfortable with that and continue on. Eventually, it will work out if you don’t give up. You always find your keys in the place you look because, well, hopefully, you stop looking once you’ve found them! But you can’t give up, and you just have to keep finding a way, trying another play.

32:17

Well said, and I guess today we call it just pivoting - right! Try another play.

32:23

You’re right! I don’t know. I feel that BCs think that’s a bad word, so I try not to use that word. It seems to make everybody scared.

32:33

All right, there you go, Phoenix! Finally, I always like to close out these conversations with understanding how you find your personal inspiration, especially being the loneliest person in the Bay Area, I guess at this point!

32:49

I am a veracious audible reader, and my playlist is everything from organizational development and people management… my undergrad, actually, funnily enough, is in Psych, so I love understanding humans, brains and how they work, technology, organizational structure, and go-to-market forces, so I listen to a million books on audible, any chance I get when I’m doing anything that doesn’t require and a kind of passive listening. That’s what I’d say, probably my number one place that I look for help sorting through, making choices, and understanding how to approach things.

One of the people at Momenta has become a great peer and confidant, so finding people that you trust that have the same values has been really helpful for me just to have some kind of tribe, I’d say, is the other half of my inspiration. Find smart people who will look at the world in a really smart way, be willing to take feedback and be wrong, listen, and do a lot more listening than talking. I just try and really pay attention. It’s funny because, before Nubix, which is now all sensor-based, I used to make the joke that my antenna goes all the way up. I definitely feel that way now, I’ve got a lot of sensors, and I am always looking for the signal that helps me navigate the next step in the journey.

34:47

Well said, I really like that. However, I’ve got images of the old My Favorite Martian, with antennas popping up, whoop!

34:56

And the little static electricity signal is going between the two, yes.

35:01

Absolutely, absolutely. Now we’ve just dated both of ourselves! This was pre-Netflix for all of your younger generations.

Well, Rachel, thank you for spending this time with us today.

35:16

Thank you, it’s been awesome, I’ve really enjoyed it. We should have done it sooner, Ken!

35:20

We absolutely should have, you’ve put so many notable Quotables in the first couple of questions there that I think there’s plenty of room and opportunity to deep-dive into some of those in future talks. I greatly appreciate you taking the time.

So, this has been Rachel Taylor, CEO of Nubix, and if I can get away with it, aka the Phoenix, because I think it’s absolutely a good description for you.

Thank you for listening, and please join us next week for our next Momenta Digital Thread Podcast. Thank you, and have a great day.

You’ve been listening to the Momenta Digital Thread podcast series. We hope you’ve enjoyed the discussion, and as always, we welcome your comments and suggestions. Please check our website at momenta. one for archive versions of podcasts, as well as resources to help with your digital industry journey.

Thank you for listening.

[End]