Ron Teeter - VP of Engineering at Jobvite

(Joel Beasley at 00:00:00) Hello, my friends. Today we are talking to Ron, the VP of Engineering at Jobvite, and we discuss the importance of building a network before you need it, making the transition from doing things to motivating others, and at what point an organization should think about hiring a chief architect. All of this right here, right now on the Modern CTO Podcast. Here we go. This is the Modern CTO Podcast.

(Ron at 00:00:35) How are you doing, Joel?

(Joel Beasley at 00:00:37) I'm good. I'm actually super excited to talk with you. I saw in the past, in college I think it was, you got a degree in aeronautics and astronautics. Is that true?

(Ron at 00:00:48) Yeah, that was actually my first part of my career, working on the space shuttle program with Rockwell. So I had two great loves growing up as a child of the sixties: watching the moon landing and then seeing the advent of computing. And I wanted to pursue both of those, and I found a way to do that. I got my aerospace degree, worked on the space shuttle program, did a lot of programming to support the design analysis, and then ultimately just pivoted back into commercial software during the nineties and caught the wave on the startups during the boom. And it's just been one startup after another since then.

(Joel Beasley at 00:01:30) Oh, that's amazing. That sounds like a lot of fun. Why did you decide to leave the space area? Was it because of that time when NASA wasn't doing a whole lot before SpaceX and Musk came on the scene and they were like, let's own this?

(Ron at 00:01:45) Well, yeah, ultimately that was a big part of it. During 1995, I was part of the commercial space transportation study that was an across-the-industry group of thought leaders that came together and said, what do we need to do to commercialize space? And we developed a lot of necessary technologies that would enable that. And it was pretty clear right away that those technologies are twenty years out. So we're not going to do any of this now, so why don't I go do something different? And it's been really exciting and rewarding to see everything that the new business-led companies can do when you're not constrained by government funding.

(Joel Beasley at 00:02:32) Oh yeah. I'm excited. We're having another person on, I think next week, called Firefly. And the guy had worked at SpaceX and worked at Jeff Bezos' company, Blue Origin. It seems like if you look at their LinkedIns, they go from one to the other to the other, and I'm like, recruiters are just making a lot of money.

(Ron at 00:02:50) Well, yeah, in a lot of ways it's the same group of people. It's such a small, close-knit community. People are going where the work is exciting, and certainly looking forward to seeing what transpires with the Starship program as well now.

(Joel Beasley at 00:03:08) Are you seeing... I mean, you work at a company, Jobvite, and I'm assuming you guys do something with applicant tracking and jobs and things like that. But do you get data on different trends, different industries that are moving? Do you guys look at that type of stuff?

(Ron at 00:03:23) Oh, of course we do. In fact, SpaceX is one of our customers. And so what we do is we build tools and services that allow our customers to seek out, engage with, get people to apply, and qualify and hire candidates so that they can build great teams. And this year's obviously been a very challenging year for a lot of companies on the hiring front. It's been interesting to see which companies are still hiring aggressively. Logistics and transportation, gig economy companies are still hiring really hot. Companies that are not: restaurants, bars, retail, exactly what you'd expect. But coming out of the summer, we saw that start to pivot and turn. And coming into September, it's like, alright, we're right back at where we were in February. And now again, we're hit with all of the more restrictive lockdown decrees to control the ICU capacity. And so we can imagine it's going to take another undulation before it starts on the road to a permanent recovery. But there's a lot of excitement in the industry about getting back to work. It has been interesting to see how much new volume we've been getting. As soon as it started to look like it was going to recover, people were on the hiring fast.

(Joel Beasley at 00:04:53) So I'm curious, where are you located physically?

(Ron at 00:04:57) I'm physically in the Bay Area. We have a facility in San Mateo, but Jobvite is a global company, and we're all working remotely now. We still have a few physical facilities that we use for postal drops and stuff. But generally, everybody's working out of their home. And it's been an interesting journey, I think, trying to understand what works, what doesn't work, and what do we need to do more of to compensate for that lack of physical presence.

(Joel Beasley at 00:05:32) Yeah. Because you mentioned a minute ago you said that with the restrictions coming back on, and where I am, the restrictions haven't really changed.

(Ron at 00:05:41) Yeah. This is a California thing.

(Joel Beasley at 00:05:44) Oh, okay. Got it. So they're being more restrictive right now. What's the... It's December. It's the first couple weeks of December right now, and they're being more restrictive than they were two or three months ago?

(Ron at 00:05:55) We've gone backwards. So where we were located, we were coming out of a more restrictive tier into a less restrictive tier. California has a four-tier system, and we were moving forward, which was great. And the problem was, I think, people just kind of overdid it. And now that caused a spike in occurrences, and now they're taking a step backward. But they didn't step backward one step. They stepped backward two steps. And so now it's kind of back where we were almost in April. So it's not... We didn't back up to July. We backed up to April. And hopefully, we'll only need to stay like this until January 4th, but it kind of puts a damper on the momentum that we had coming out of September.

(Joel Beasley at 00:06:47) That's great. I don't watch the news. So this is news from right now. You are my weather person for the coronavirus weather. I don't know what to do with this right now. I had no idea that all my friends out in California, they're locked down again right now. Wow. Things are definitely Wild West down here. It's amazing to see how the different states handle things differently. It reminds me that when we hear about other countries in other parts of the world, if you look on a map, we are a bunch of little countries, right?

(Ron at 00:07:17) Well, that is the United States, right? The states are sovereign. The federal coalition is the union, and this is why states' rights are so important. It should be up to local people to decide how they want to govern themselves. But at a larger level, there is a role for federalism as well to say, yeah, this would be a good idea for everybody to do because we're all in this together and because our borders are incredibly fluid between the states. What happens in one state affects what happens in another state. So I think there's a balanced measure here, but it's pretty clear that there is some benefit to wearing masks. And generally, people should do it, not necessarily because it protects them, but it protects the people around them.

(Joel Beasley at 00:08:08) What I like is that, or the reason why I don't watch the media, is they make it seem like everybody's so angry. And then everybody I talk to is completely reasonable. It's like, yeah, we're all American. We love the core values of America. And so I'm on the team of, I think right now, as far as people, as a civilization goes—and we'll get right back to the technology stuff—but I just think the unity right now is the phase we need to be in. We need to come together and figure out what do we have in common, and how can we just...

(Ron at 00:08:41) Yeah. But Kumbaya doesn't sell, right? The media companies generally are targeting toward the outrage cycle. They need the clicks. They need the ad generation. So you have to look at how the product is packaged to understand how it's going to be consumed. It's not in their interest to say, hey, everything's fine except for these wackos on the extremes.

(Joel Beasley at 00:09:04) But I disagree a little bit with that because the thing that makes this media outlet successful is just being useful to other people, right? So there is another model. I mean, there is another model. Their old models—and that's the problem. They had to transition old models. Yeah.

(Ron at 00:09:23) Yeah. And you're right. You've got things like Modern CTO and the TED talk tracks that are generally informative and trying to improve the situation for humanity. But you also have these other revenue-generating publications that are trying to get some news out, but they're trying to spark a debate, and they tend to amplify extremism rather than amplifying that... That's not the bell curve. The bell curve is here.

(Joel Beasley at 00:09:57) Yeah. It's like when I was young, when I was going to the grocery store when I was eight or ten, on the checkout line, there was all the junk, the junk publications, right, the nonsense. And so that was a tiny little fraction, maybe let's do the Pareto thing, that was 20%. And then you go home and you watch the news and it's kind of boring and they're doing these extensive reports and investigations, and that's the 80%. Seems like we live in the world that's flipped now.

(Ron at 00:10:25) Yeah. Well, to use your analogy, there used to be the aisle in the store that says, here's where all the magazines and books are. And then there's the stuff that's right next to the checkout counter. And it is the can I get some of your money from an impulse buy? Because you wouldn't go in and buy this off the shelf. It's just because it's right there. It's all about placement. And in a lot of ways, the digital landscape that we live in has created the optimum placement because I can put whatever everywhere. And so then it's what are people consuming? It kind of depends on the impulse click.

(Joel Beasley at 00:11:10) Oh, that's good stuff. I want to talk more about... I will talk about that stuff forever and get way off topic, so I apologize for that. But I want...

(Ron at 00:11:17) No worries.

(Joel Beasley at 00:11:18) I know Jobvite does a number of things, the applicant tracking, all sorts. Can you give me your 10,000-foot overview of Jobvite?

(Ron at 00:11:27) I think the easiest thing that you can think of is if you had a factory and you're trying to manufacture things, you've got a process. You're pulling things through the factory. You want to optimize that process. You want to make it fast, frictionless, and inexpensive. We bring a set of software and services to the hiring process to do exactly that. Trying to find qualified candidates is probably the most expensive part of the whole process. People are posting jobs out on job boards. They're posting through social media. We automate all of that infrastructure as well as try to report on the effectiveness of those placements so that people know the value of where they're spending their money. Are they getting the quality candidates that they need? And then once they're getting them into that pipeline through an apply, how are they doing through the qualification and hiring process? It's a complex process. And if you've ever hired somebody, you know there's a lot that's going on. Imagine if somebody's trying to hire 2,000 people a week. You need a different kind of automation and a different kind of scale to support people with those kinds of demands. It's easy to hire one or two people. But if you're doing six a week as a single hiring manager, it's already more complex than what you're going to be happy with using an Excel spreadsheet with. And that's really where the applicant tracking system came in. We do recruitment marketing to handle the front side. It's flipping around rather than having you passively come to my job. How do I get to you and show you that my job is there? That's really the recruitment marketing piece. We've got excellent technologies and services there to help accelerate that outreach. Social media is what Jobvite's known for. Jobvite is a social media job invite. And we came through that in the prior recession. 2008 through 2011 is when we built out a lot of our functionality so that we could grow out of that recession into the hiring phase that came afterwards. And we're doing something very similar this year as well. We're investing a lot of new features and functions this year so that as people start hiring and ramping out of this, they're going to be even more effective next year.

(Joel Beasley at 00:14:01) That's smart. Use the time appropriately. Are you going through M&A? Are you building your own products 100% in-house?

(Ron at 00:14:10) We've done a lot of M&A over the last few years. Jobvite's moving out of that teenage to young adult phase in a company. So two years ago, we got a majority shareholder. We went out, we bought some companies that were partners and good matches for us in terms of having features and functions that operated at enterprise scale in a way that it was going to take us time to build. But since then, we've been focused on integrating these things and making them work really well together. So our Talemetry, RolePoint, Canvas, and Jobvite product lines are all now integrated. We have a common login system. We're working on building out the best suite in the industry. And it's going to take us a little while to get to that vision. But the fact that we're investing now while things are slow allows us to be able to accelerate out of this when the hiring returns later next year.

(Joel Beasley at 00:15:14) So in my town, when I was raising capital, there was this other company that was in the HR space. And they were a small startup. I think they had five, seven people, but they were doing this thing that I thought was pretty fascinating, where they were understanding the psychological profile of the type of team member that you needed next. And then they were scanning the Internet and building millions of profiles of potential candidates and then analyzing them against that and connecting them. Is that something that's regularly happening in your industry, or is that something that was just news to you?

(Ron at 00:15:53) From a psych profile, probably not highly common. What most of the industry is focused on right now is really just trying to provide more intelligent screening. How do I find people that are a better match to this job? So that's the how do I go out and source people to match the job that I want? And then the other part of that is how do I get better at qualifying people? Because those are the two things that tend to take a lot of time. We've been investing in things that help the candidate and the recruiter communicate. So moving from an email-based messaging system to a text-based messaging system, bringing advanced chatbot technology into the product so that some of the easy questions can be offloaded, and really focusing on how do we maximize the recruiter's time in terms of generating successful outcome for the company. It's not always a successful outcome for the candidate. Recruiting is not like sales. We're not trying to drive as many candidates to the success goal.

(Ron at 00:17:06) You're trying to drive the best candidates to the success goal without giving the other candidates a bad experience, because if you have a secondary position, you want to take your silver medalist from your first job and try to reroute them potentially. Right? And so focusing on candidate experience in a process that's generally really negative, because, you know, let's face it, most people are rejected from most job applications.

(Ron at 00:17:32) It's kind of the nature of the industry. But it's certainly better to get a nice rejection and some feedback than, you know, your application going into a black hole and you never hear from the company again.

(Joel Beasley at 00:17:45) I know. I hear so many people—that's usually the first spot when they're figuring out how to find a job. Like, I'm applying to 80 different places and no one's responding. That's something I hear all the time.

(Joel Beasley at 00:17:57) And I'm always like, go put a sign up and make some money, like be an entrepreneur. No, I realize that's just the way I think. But it's important for people to understand that relationships are so important. Like, building up a network before you need it is what you have—it's like savings before you need it, right?

(Ron at 00:18:19) Exactly. Exactly, right. And that's part of why, you know, Jobvite pivoted into the social media broadcast mechanism very early, is because we saw that the power of your network was—first off, your network is full of people that you like, right?

(Ron at 00:18:37) There are people that you have or do like to work with, and they generally produce the best hires long term. So why would I go waste time, you know, posting my job to a public job board where anybody can come and apply to it and I have to throw away 90% of the applications, if I can get some sourcing from a network of trusted people first, right?

(Ron at 00:19:00) And so now what you see, there's a lot of focus on building trust networks a priori, so that when you're searching for a job, you don't have to go through the finding and qualifying these candidates. They're already prequalified. And you can think of, you know, a lot of the hiring agency type use cases are—well, they'll go find and qualify candidates for you so that it can shorten your cycle time on getting your position filled.

(Joel Beasley at 00:19:29) Yeah, which is important because we learned that when we put job postings out. We have about 10 people here at the podcast, but, you know, we're hiring a person like every other month now or so. And 80, 90% of the stuff we get, I've just like—I don't see why this person is applying. Like, they don't have any interest in this in their life.

(Joel Beasley at 00:19:48) And then I remember back to the earlier days when you're trying to figure out what you're doing and who you are and kind of exploring like a child, right? Like walking around the house playing with everything. And then I relax a little bit, you know?

(Ron at 00:20:06) Yeah. And I'll say that, you know, there are people who like to do more of the same, right? And so they'll bring their expertise to the job. And then there are people like me who pivoted out of, you know, aerospace engineering into computer science full time.

(Ron at 00:20:24) When I came across, it's like, yeah, but you've got an aerospace degree. And I'm like, yeah, I know that. But, you know, I really enjoy programming, and here are some examples of the programs that I've written, you know.

(Ron at 00:20:35) And it made that transition really easy, but only after I had an opportunity to have the conversation.

(Joel Beasley at 00:20:43) You should have been like, yeah, I'm really great at launching products or projects, right? Like, you should have thrown in some wordplay with the aeronautics.

(Joel Beasley at 00:20:51) With the astronomy. Be like, yeah, the stars are aligning. That's why we're here today.

(Ron at 00:20:55) If you're looking for a rocket scientist, I am your guy.

(Joel Beasley at 00:20:59) That's awesome. You are so cool, Ron. This is fantastic. I'm so—I'm going to thank Leron. Like, I'm going to be like, dude, thank you so much for introducing us, because that's how we got connected, through Rookout.

(Joel Beasley at 00:21:09) How did you end up meeting them?

(Ron at 00:21:13) We actually met the Rookout team because we have a relationship with Cisco AppDynamics. We've been an AppDynamics customer since 2013. And one of the things that I like about Jobvite is that we don't treat vendors as vendors. We treat vendors as partners. And we build these really deep partnerships with them so that we're cross-collaborating, you know, sharing feedback, pre-evaluating new features and functions.

(Ron at 00:21:46) And when you do that over a period of years, it builds a relationship where they're saying, hey, you know, we've got something new, and we want you to take a look at it. You know, we were one of the first companies that got to see what Rookout could do, and it was an amazing and immediate win for me. I mean, to me, it took me about 30 seconds to think through everything this was going to do for us. And I'm like, how does this not sell itself?

(Ron at 00:22:12) Right? Because essentially you're looking at a tool that allows you to do production-time debugging without stopping production-time traffic, right? So it works just the same as a debugger would on a local development environment, except it doesn't break the thread. And it's just—it's beautiful. It's elegant. It's simple. And, you know, we pulled it into our infrastructure almost immediately. It's been a huge time saver for my team.

(Joel Beasley at 00:22:43) Yeah. I was talking with Jason, CTO over at GitHub, about it. And then right before, like, five minutes before the show was going to start, I quickly realized that the CEO of GitHub was actually invested in Rookout. And I was like, oh, wow. So Jason's going to know who Rookout is. And he definitely did.

(Joel Beasley at 00:23:02) It was such a cool moment. And then I thought at that point, I was like, this is—I mean, that company is going to just take off and explode. Then I met Leron. He was such a cool guy, and then they had—as an engineer, when I first saw their product, I was like, this is so neat.

(Joel Beasley at 00:23:17) Because, you know, you just think about all the—I'd say two weeks before I found their product, I was having a conversation at a conference about anonymizing production data down for development, right? And I was like, oh, this is so neat. I can't wait for that conference to come back next year. I would go talk to those people about this again, because I need to check out Rookout, you know?

(Ron at 00:23:40) Yeah. And that was one of the things, you know—part of the benefit of being early is we had an opportunity to provide a lot of feedback to the team. And they were amazing partners in listening to that feedback. So we were able to show them a path that worked better for us, and they were able to incorporate that in the product. But if you think about some of the bigger things like zero trust, I want to have a zero trust posture around my production systems.

(Ron at 00:24:12) So how do I get engineering the data they need to reproduce an issue when it's happening in production and I've locked them out, right? Rookout is your saving grace, right? Because they can go in and they can get the data they need without ever being exposed to the data they shouldn't have, right? And it's without having any access to any of the production infrastructure databases. It's just literally in the scope of this code, here are the variables that are in scope. And these ones are redacted. So, you know, you can't have these, so you won't get them.

(Ron at 00:24:45) It's just elegant and beautiful again.

(Joel Beasley at 00:24:48) When you have to sell an expense like that to the team or budget for that, and you have to kind of figure out how to think about it in the context of the P&L, did you go for like security cost reduction, or did you go for efficiency and time of engineering? Like, where did you—how did you quantify this as an investment?

(Ron at 00:25:09) For us, it was really around customer outcome. You know, we have SLAs on turning around customer issues as they're brought in, and we would like to make our SLAs tighter. We want to be the best service provider in the industry. And in order to do that, I've got to take a process that takes hours or days and shrink it into minutes. And Rookout gave us that ability.

(Ron at 00:25:35) So it was an easy win because I've got support from our senior leadership on that kind of a priority item. You know, there are other reasons to do it as well—it's a more efficient use of our engineers' time, ergo, they can deliver more meaningful product as well. But the primary purpose was to shorten the customer impact time. You know, from a DevOps or a DORA perspective, it's the mean time to resolution, right?

(Ron at 00:26:03) By reducing that by 80% has some meaningful outcome on customer goodwill and retention. So it's generally pretty well an easier sell. But there are other things you can do with it. It's like, I can run it in my CI environment, and I can get information upstream in the cycle. If I'm running a shift-left program, I could run Rookout locally, and then I don't even have to worry about whether or not, you know, I can get my IDE to connect to my running process in Kubernetes. Who cares, right?

(Joel Beasley at 00:26:38) Now, you have like 100-plus engineers. They're like my guesstimate off of that LinkedIn data, right? And I don't know how accurate that is, but how do you go about selling this or deploying this type of product throughout your engineering teams? Like, how did you do that?

(Ron at 00:26:56) It was interesting too, right? So we deployed it for a proof of concept. We showed it to a couple of people. And at the time, we didn't have single sign-on, so we were having to provision accounts for specific users.

(Ron at 00:27:10) And what was interesting is to see the backup in the access requests as engineers were hearing word of mouth from other engineers who had access about this. They wanted it too. And, you know, because we didn't have SSO, they had to ask for an account. And just watching it kind of, you know, almost double every day. It's like two, four, six, eight. You know, pretty soon the whole team was going, when can I have access?

(Ron at 00:27:38) For us, there's a technical limitation for us as well. Rookout is an agent that runs in your primary language. So, you know, it was very easy for us to adopt the Java agent. There's a C# agent, and now we're prototyping a Ruby agent. A big part of our infrastructure is Ruby, and there's no agent.

(Ron at 00:28:00) So it's like, although we wanted the capability, it wasn't there. And again, you know, that's where the Rookout team really heard our urgency around Ruby, and they prioritized getting the agent developed and ready for us for preview.

(Joel Beasley at 00:28:15) Nice. I'm a fan. Ruby is the language that I program in most recently. So about seven or eight years of Ruby. It's a—once I started using it, it was just such a beautiful language to use. And then it had, of course, like, the Rails framework and all of that.

(Ron at 00:28:33) Yeah. Yeah. It's nice when you have, you know, those good archetypes built into the framework so that you don't have to keep reinventing the same solutions mechanically. You know, we use a lot of Java Spring, and we use Ruby Rails. We also have a lot of Python that we use for either serverless or, you know, small use cases. But, you know, for us, the big transition now is moving from, you know, EC2 hosting into the containerized versions or going fully serverless for some other items.

(Joel Beasley at 00:29:10) Yeah. I remember when I first found out, like, frameworks were a thing, because when I started, I didn't really know that they had all of these collections of tools that you could use. I was kind of just writing my own stuff, and I was like, I'm so cool. I've got this really amazing asset. I reuse it on multiple projects. And then I came across—my buddy Derek's like, hey, let's check out like CodeIgniter or some framework like that. And I was like, this is unbelievable. And then I just found Ruby and a couple other ones and CakePHP, all these different frameworks. And I just was like, this is—it just—it took—you would need a team of 20 people to do it. And now you could do it with one person in an afternoon. Build a basic app.

(Ron at 00:29:51) It's amazing how many great frameworks are out there. And I mean, part of the challenge is there's maybe too many, you know. And I don't want to be heretical there. But, you know, having some really, really good frameworks is better than having a lot of okay frameworks, right?

(Ron at 00:30:11) And to me, you know, it continues to baffle me that, yes, there's these great frameworks for writing applications, but there's no great frameworks for writing business applications, right? There's another level above that says, here are all the layers of business requirements that sit on top of, you know, computer science coolness, right? It's like, these are the kinds of things business users need regardless of whether you're doing a payroll system or whether you're doing an ATS or whether you're doing, you know, a Ticketmaster, right?

(Ron at 00:30:43) There's this set of business requirements that I have not seen change in the last 25 years, and there's no frameworks to deal with that, right? People are providing language or they're providing technology, but very few people are really addressing that business solution layer. You know, obviously you can build things in Force.com to do that, or, you know, but that's not really portable.

(Ron at 00:31:10) And so, you know, people have to pick and choose very carefully. If I'm going to go that route, then I'm stuck in that ecosystem. If I want to do something different, then I'm kind of on my own, right? So, you know, there are definitely ways to get it done, but it just—to me, it seems like, you know, there are thousands of companies that are resolving the same problem over and over again. First name, last name. First name and last name. Last name, comma, first name, right? It's like—

(Joel Beasley at 00:31:39) Oh, yeah. It's amazing the amount of duplication. I created an analogy to somebody in the leadership development world where what I saw happening with companies trying to recreate WordPress versus just using WordPress, and it's just like, you don't need to recreate it. You don't need to spend all that engineering power. Just use something that already exists to manage your content, right?

(Joel Beasley at 00:32:02) And then it's open. So then start customizing it, versus, you know, them creating these leadership development programs from scratch constantly. And so I was like, oh, man. This is amazing. It's the same mistake I see over here, but it's happening in this world over here because they didn't make this mistake.

(Joel Beasley at 00:32:18) And engineers have since grown and learned this mistake. But this part of the industry over here and this market has not. And it's like, oh, is this what experience feels like, right?

(Ron at 00:32:30) Yeah. Well, that's exactly the case. And if you think about it, think about that same kind of mentality applied toward hosting infrastructure, right? You know, we started with on-prem, and you would go buy network gear and lease power and high back. And you'd get a rack, and you'd rack servers, and you'd build a drive, and you'd network it together, and you'd apply firmware, and then you could start deploying stuff, right?

(Ron at 00:32:48) Right. And that's really where, you know, the cloud comes in, and it's that framework for hosting that takes a lot of that away. But the mistake I see is people go into services like AWS, and they provision the network, and they provision the storage, and they build hosts, and they get it all set up, and then they can deploy.

(Ron at 00:33:20) You know, it's like you're just repeating the same set of mistakes because you're unwilling to engage fully with what the cloud can offer, right? It's like, I just want my method to run. Why am I not using Lambda, right? Why do I need an EC2 host to run this? You know, it's that data center in the cloud versus actually building a cloud application.

(Joel Beasley at 00:33:43) So your team must be pretty excited to work at Jobvite. Must be pretty advanced then, with you've got such a long history of experience as a chief architect. I always try to figure out, like, where's the area that this person, you know, where are they gonna really shine and have a lot of experience? And I just saw throughout your history, you're like Mister Chief Architect.

(Ron at 00:34:03) I've earned it, you know. I would say I choose difficult situations. Right? I'm not interested in joining a startup that's, you know, pie in the sky and clean slate. I usually join, you know, round A, round B companies that, you know, they've had some initial success, but they're running into some scaling challenges because that's what I find. It's that crunchiness of how do you make this scale on a budget that I find really interesting and challenging.

(Ron at 00:34:37) And having done that a couple of times, you know, Jobvite's in a really good spot, and now we're able to accelerate because we've got control over that. We have a very competent cloud hosting team, and we're moving that skill set out into our application development teams so they can own their own delivery cycles. You know, that's the ultimate goal of what we ostensibly call DevOps. But what we've been finding is that DevOps is actually too small. And so people start talking about DevSecOps because, obviously, I need it to be secure.

(Ron at 00:35:14) And that's interesting also, but it's kind of the wrong order because the security requirements should come first. Right? And then the other observation that we've made is there's a stakeholder that's not at the table that needs to be, especially in SaaS, and that's finance. It's like, when you're talking about, hey, I'm gonna go deploy this new feature, it's kinda like, well, what's the budget for that? And who's paying for it? And how are you gonna get that budget in so the finance knows that spend is coming in February? Right? So that you can help them manage cash flow. Right?

(Ron at 00:35:48) So what we're actually kind of prototyping is something I call FinSec DevOps. Right? Get the finance and security requirements first and then do the DevOps work.

(Joel Beasley at 00:35:58) So is that like an internal process, the way you have things structured at Jobvite?

(Ron at 00:36:04) It is the methodology that we follow. Right? Where's the budget? What are the security requirements? What are the feature requirements? What are the operational requirements? Go build it and operate it. And as we move forward, the application teams that build the software are gonna be responsible for operating the software. Right? There's no motion of that to some other team. There's no SRE team. There's no operations team. It's you are responsible for everything you deploy. And so if you deploy something poorly, that's going to interfere with your next sprint of planned work because you're gonna have to respond to that. Right?

(Ron at 00:36:45) And engendering that level of accountability and ownership, I think, over time, will help the teams deliver superior products. Because, you know, the bigger problem that people have is a lot of the teams that write code don't know how code actually runs and operates at scale. They just write code. And so they keep making the same mistake over and over that some other team has to try and fix, and then you get into this communication flame war. Right? This is what DevOps was intended to fix. But it only works if you actually follow it all the way through and say, no, this is your infrastructure. Your team is responsible for the spend as well. Right?

(Ron at 00:37:28) Not just the availability, not just the meantime to acknowledge, not just the meantime to resolve, but you're also accountable for the spend. So if you double your costs, the business is gonna come back to you and say, why is this happening? Is it because of usage? And how did you think somebody was gonna pay for that if it's not billed? Right? The goal here is to make a functional business that can scale.

(Joel Beasley at 00:37:53) Yeah. And I like that you mentioned ownership because that's, like, step one. You have to have a team that has that culture of ownership in order to grow something really big and scale it.

(Ron at 00:38:05) Yeah. Yeah. And interestingly enough, your ownership comes through vulnerability and transparency. Right? And those two lessons are hard lessons for teams to learn, especially, you know, depending on where people came up through their career. You know, being transparent and being vulnerable isn't easy. But you can't really get to accountability and ownership without passing through that filter first. And, you know, trying to do that with 13 scrum teams, you know, is a little bit challenging. Some teams get it faster than other teams do. But, you know, as an overarching organization, we need to continue to keep pacing forward.

(Joel Beasley at 00:38:46) So what makes, with all your experience as an architect, what makes a good chief architect?

(Ron at 00:38:52) Oh, boy. That is probably the hardest question anybody's ever asked me. You have to be willing to listen and you have to know how to share experience without overwhelming the conversation. The key is to try and show people the path that you see when they're still looking at their feet. And you're, you know, maybe up at the peak and you're looking across to the next valley. You're trying to communicate to them not only where you've already been, that they're still progressing, but where they're going to go next. Right? And sometimes that gulf in experience is very difficult to bridge. And so you have to have a tremendous amount of patience of understanding where people are and helping them see that there's a path and then trying to work to accelerate their path so it doesn't take them ten years to do the same transition. Right?

(Ron at 00:39:53) You know, the business doesn't have time for people to learn everything the hard way. And so you have to build an open culture where people are unafraid to ask questions, but also not afraid to experiment and fail. It's one of the things that, you know, I think significantly differentiates companies like SpaceX from NASA is that risk versus risk-averse behavior. Right? SpaceX can, because the shareholders are the only ones that care if they fail. Right? If NASA fails, then Congress could cut their budgets. Right? So there's just an institutional resistance to risk. But, you know, you have to take risks in order to learn. You know, the lessons that we learn without failure are probably the least important lessons we can learn in life.

(Joel Beasley at 00:40:52) Yes. A hundred percent. The lessons that cause me the most pain are the ones that I remember for the longest. They're the ones that just you don't forget those, and they become the most important lessons of your life in hindsight.

(Ron at 00:41:10) That's frequently the case. Yeah. You know, to continue on that thread, I would say, you know, in a highly evolving technical world, one of the challenges that I have had in the past is just maintaining an awareness of where the state of the art is. You know? I think we're really good right now, but, you know, maybe, you know, eight, nine years ago, I was so focused on the problem that the company was trying to solve at my prior employer that I was losing sight of what was going on outside of me, outside of the industry. Right? And part of what I'm working on for my personal improvement is doing things like this, where it's like, I have to be more engaged with what's going on outside of the company in order to really take what I can do inside the company to the next level. Right? You have to be out and collaborating, not just consuming passively, but you also have to be out and sharing.

(Joel Beasley at 00:42:12) Yeah. So when, when we have a lot of companies that are in the growth stage. Right? They're maybe, let's say, they've got 25 engineers or so. I mean, between, like, five and 25 engineers, and they're growing and growing, and they're excited. At what point? Like, what's the flag or the thing that would, like, tickle their spidey sense that would let them know they need a chief architect versus just kinda figuring things out?

(Ron at 00:42:43) It really comes when you start to realize, you know, it kinda depends on how you started. Right? Jobvite started with a monolithic application server design. It's great for prototyping, very 2005. You know, that's kind of the way things were being done then. But by 2008, it was already showing some challenges in scale. Right? You can only vertically scale so large, especially if you're on-prem. You imagine, oh, I need more memory in my server, but my server is maxed out. So now I need to go buy a new server and wait for it to get shipped so that I can install memory, and then I can move the software. Right? It's very difficult. Right? Companies that are born in the cloud generally are going to be more nimble and more safe from some of this, especially if they've made good choices on their design patterns for the cloud. Companies that started on-prem or started in the cloud with data center in the cloud as a mentality, it really starts getting into scale.

(Ron at 00:43:50) You know, you're going to start seeing complexity or scale dominate the discussion. And at those points, you need some leadership that says, okay, I've seen this. I've done this before. Here are our options. Right? Let's collaborate with the team, and let's decide what we're gonna do. Are we gonna refactor this monolith into microservices? Are we gonna host it on Kubernetes? Are we going to pivot into serverless? What are we doing about databases? What are we doing about search? Right? What are we doing about cache? Right? Those are the things that obviously contribute to a system's performance as it scales. But then you also get into that, you can throw money at a lot of these problems, but at some point, it's like, you know, when the money starts, when the money curve starts to flatten, right, they always do. Right? It accelerates up and then it inflects and it flattens again.

(Ron at 00:44:46) As the money curve starts to flatten and your costs are going up, that's when you need this other perspective to come in and say, how am I going to pay for this profitably? Maybe I need a different design that utilizes cheaper storage or cheaper compute. Right? And that's kind of the driver that we're focusing on right now is we can manage the deployment cycle and provision new services very easily. How do we do it cost effectively? And how do we do it cost effectively based on people time as well as compute time? And, you know, some of our mantras is, you know, we prefer services over servers. Right? If we can find an AWS service that does the same thing, we will always choose that over running the service ourselves. Right? Whether that's a database or search engine or a streaming event queue, there are, you have to recognize where the expertise lies and then just say, this is either core competency or it's not.

(Joel Beasley at 00:45:53) Yeah. Because they have people just eating, breathing, and sleeping that specific service and keeping it up all day. Right? Yep. Leverage that. So what are some of the, like, interesting, difficult engineering problems that are specific to the recruitment industry, if there are any?

(Ron at 00:46:15) Well, there are two, and I'm sure my chief data scientist is gonna kick me later. So I read an interesting statistic the other day that said that 78% of people lie on their resumes. But if you're using their resumes to do the job matching, then how effective is the job matching going to be? Right? So you have this challenge in recruiting that you've got humans on both sides. Right? You've got a job description that is marketing speak for the job, but it's not probably really what the job is when you get it. Right? And then on the other side, you've got somebody who's marketing speak for themselves and their skills, but it's probably not really what the person is when you get them on board. Right? So you're trying to match these two flowery descriptions of what's real. But on the back end side, you've gotta deal with the reality of it. Right? And so, you know, for us, it's like, how do we build a more accurate predictor of compatibility based on the fact that we're ingesting these pieces of information that may or may not be accurate.

(Ron at 00:47:29) Right? And I think we came up with a really amazing way to say, well, I can tell how much fuzz is in this resume because the things that they're listing in the keywords are unrelated. Right? So versus having a strong correlation. Yes. Somebody who was a chief scientist would know R and databases and this, but they probably don't know React, right, which is a front-end technology. So, you know, having a model that can tease out these things are true and those things are unlikely to be true allows us to then exclude those kinds of things from primary match. And we think it gives us a better match than a lot of the, you know, kind of brute force approaches that are being taken in the industry.

(Joel Beasley at 00:48:23) That's interesting. Now do you, like, pick certain cases or do you have, like, teams that, like, follow through with some of these to actually figure out if these assumptions, like, how valid they are? Or how do you improve it and follow, like, when you have this assumption about, like, they might not, this thing's unlikely to be true. Right? How do you better understand that through real-world experience?

(Ron at 00:48:48) So part of that is to, you know, build an engagement in the product that allows the recruiters who are evaluating these candidates to say, yeah, I agree with that. No, I didn't agree with that. Right? To start layering in their own preferences or biases. But one of the nice things about being Jobvite is we use our own product to do our own hiring. Right? And so we can see it in action, and it's kinda like, why is it that that guy didn't get a good score even though he's the one we picked to hire? Let's go begin and figure that out. You know? So we've got real-world experience with doing it, but we also have features in the product that allow the customers to tune the results to their needs. And while I'm on the topic of biases, you know, this year was a big social justice year for awareness. And, you know, Jobvite's working on some diversity inclusion capabilities and features as well that we can use with the same sets of data. Right?

(Ron at 00:49:50) Is this job description biased against certain people just in the way it's phrased? Right? I think a useful feature to help hiring managers be more aware of their own biases, that hiring managers or HR, whoever's authoring the job description, can be more aware of the verbiage that they're using and whether it's going to automatically self-exclude people.

(Joel Beasley at 00:50:18) That's interesting. So I didn't know about this. So people writing job descriptions, you can have bias in the job description?

(Ron at 00:50:27) Oh, yeah.

(Joel Beasley at 00:50:28) Oh, okay.

(Ron at 00:50:28) Hey, I'm looking for a rockstar. That's not always what people are looking for. And the affinity for ninja, right? Some of these terms are automatically exclusive based on gender or based on regional identity. And so, you know, making people more aware of their subconscious or, you know, even conscious biases, I think is a good thing. You know, they can still choose to say, nope, this is what I want because I'm looking for a rockstar. Okay. But you're automatically excluding these populations.

(Joel Beasley at 00:51:06) But what are they? Because I thought a rockstar, I thought people, when they use that, they just mean you're gonna work really hard.

(Ron at 00:51:12) Yeah. Well, that's the thing is, like, it means different things to different people. Right? And while I don't have, you know, specific, you know, population data to share with you, it's that there are terms that you can use even in normal writing that are exclusive to females, that are exclusive to people of color. Right? It's very subtle. Right? And one of the things that, you know, we have to recognize as, you know, people with Caucasian backgrounds, you know, we have privilege. And that means that the bias is invisible to us in some cases. And that's the beauty of having this data system there is to show people that it's there.

(Ron at 00:51:55) You can see it because there's data to back it up. Now that there's data that's backing it up and you're aware of it, what do you want to do about it? Right? One of the things that we've been trying to highlight was, you know, you can't make change using AI by training it with historical data, because if you're using historical data where the bias was present, it's just going to do more of it.

(Ron at 00:52:20) Right? So you have to be aware of what bias is in your training data so that you can change the model so that it's not biased, or at least the biases are overt and you're aware of them.

(Joel Beasley at 00:52:31) I think one of the interesting things I had heard was there was some sort of hand gesture or something that in America was normal, but in another country was incredibly rude. Right? And so it's interesting because this technology is spanning the globe. It's spanning all of humanity. So I'm fascinated by this concept, specifically with words and their meaning, or gestures and their meaning. It's like, who's going to be the world police? Like, the database or register for micro grievances or whatever they may be called? Like, how do you do that on a large scale? Or is it just something that just kind of happens in these communities and it just is going to figure itself out long term? That's probably it.

(Ron at 00:53:17) Yeah. Yeah. You know, globalization is a hard problem. Right? You know, you can say, well, okay, we globalized our product. You know, it's available in 30 languages. We can show dates in this format and that format. We can show numbers with the commas here and the periods there, but that's not really what globalization is about. Right? Globalization is about understanding that there are people that grew up with different experiences than you. And if you're going to communicate to each other, you have to be able to speak each other's language. Most of the things that we think of as globalization in the product is really product to consumer. Meaning, I grew up in the United States. I'm expecting to see numbers like this. I'm expecting to see times like this, and I read English. And somebody else is growing up in France, and they learn French, and they see numbers this way, and they see dates that way. The product is solving that problem right now. But there are very few products that are actually solving this problem, is I want to talk to you, and I speak English, and you speak French. Right?

(Ron at 00:54:26) Google Translate's there. But nobody's using it really to say, hey, actually, I want to foster communication back and forth so that you can talk to each other across languages, customs, and cultural norms. Right? And that's kind of where, you know, we start getting into the cultural norms around gestures or word usage is something that if you're unaware of it, you're unaware of it. You're going to unintentionally offend. And there's a social awkwardness to that, obviously. You know, we all hear about the foreign traveler visiting in a city. You know? It doesn't matter where you're going. It's just they stand out in the crowd because they're the only ones that are not functioning correctly in the social norm there. And the question is, in a global environment where we're working across these environments, how do you not be the traveler? Right? How do you automatically fit in with your people? And it really takes a lot of relationship building.

(Joel Beasley at 00:55:28) Yeah. No. It's definitely a whole—I love to see how technology is growing and emerging. And I think it's fascinating all these little things that are coming up, whether it's in our products or our culture. Because what it's doing is it's creating some really interesting and unique discussions. Right?

(Ron at 00:55:49) What's interesting though is if you go and you look at a city, you know, any large metropolitan area, you'll see it's very similar. Right? There's villages and neighborhoods that have a culture to them that, you know, you're native—you belong there. Right? And if you go into a different neighborhood, it has a different flavor, and, you know, the shops that are there, the people that are on the street are wearing different clothes. Right? That kind of diversity is something that, you know, we should all recognize as normal and something to be cherished because, you know, it allows us to have these differences, but yet we have this layer of cooperation that allows us to get along as well. Right? The global stage is just the next layer. It's almost like recursive. Right? It's like you can have a household to two households, to a neighborhood, to a city, to a state, to a country, to the planet, to the solar system. You know, it's going to be fascinating for me to kind of envision and watch as we start thinking about, yeah, we're going to colonize Mars. Okay. Well, how is Mars' government going to work? And how does the economy work? And what is the legal tax situation on Mars? Is it going to be the new Caribbean? You know, because there are no laws in space. So it opens us up to a brand new line of thinking about how we as humans collaborate with each other in a way that's not mutually destructive.

(Joel Beasley at 00:57:33) You're like a rocket scientist, man.

(Ron at 00:57:35) I am a rocket scientist.

(Joel Beasley at 00:57:37) This is great. One last question as we wrap up. What are you learning right now as a leader?

(Ron at 00:57:43) As a leader, one of the things I'm learning is how to transition from doing things to motivating people to do things. In order for a leader to scale, you have to be able to have the ability to delegate and the ability for your delegates to enact your will without a lot of oversight and without a lot of direct involvement. And sometimes that's easier than others. So really, you know, we've been talking about developing management skills within Jobvite recently. You know? It's like having courses for people managers to help them understand how to build relationships and how to understand where employees are, work through things that are challenging, create enthusiasm for what the company is doing. And one of the areas that we've been trying to work on most recently is trying to identify and nurture leadership. Right? Because leadership isn't about having people that you can tell to go do something. Leadership's about being able to inspire people to go do things on their own and say, I want to go there. Can you manage this? All right. And then I can just not think about that anymore. I can go work on something else. And I know that that work's going to get done. And then just kind of lightly touch base on how's it going, what are the resources that you need. Right? It's really about being the servant leader that we talk about in Agile is really, you know, there to facilitate the scale of the organization. And sometimes it works really well, and sometimes it's really challenging. And I love that challenge. It's keeping it fresh for me to be learning new skills and working through the—you know, you should always try to survive a mistake that you make because that's the only way you can learn from it. So it's kind of like, okay, what mistakes can I make, and what mistakes would I not survive? And try to avoid those, and yet keep pushing the boundaries since, you know, having the imagination to try things that other people think are not possible or not worthwhile is a key part of it as well.

(Ron at 01:00:05) Because there is a reason to do something. And if you can find the business justification to do it, it should get done. And trying to create that same sense of enthusiasm and motivation across an organization of 100 can be challenging at times because people don't get it. Right? They just—I just come in. I want to write my code. You know, it's like, yeah. But what your code is doing is actually changing other people's lives. Right? It matters. And so, you know, doing this code this way instead of that way makes a difference to the company, and it makes a difference to the customer.

(Joel Beasley at 01:00:49) Thank you so much for listening. And if you found this episode useful, please share it with a friend or colleague who you think would get value from it. And if you have topics that you'd like to hear discussed on the podcast, either add me on LinkedIn, or send me an email: [email protected]. Every time I get an email or LinkedIn message, it absolutely makes my day and inspires me to keep going.