Rails Business
Brendan Buckingham and Ryan Frisch talk about developing with Ruby on Rails and how to leverage it to build a business.
Rails Business
Ryan Castillo of Knowatoa
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
In this episode, Brendan is joined by Ryan Castillo, co-founder of Knowatoa, a platform dedicated to optimizing business visibility and accuracy across AI-driven search engines. Ryan shares insights on his background in software development, and how his journey led to the creation of Knowatoa. They discuss the origins of Knowatoa, and Ryan also shares experiences with AI tools like ChatGPT and Cursor that have been integral in their development processes.
00:00 Introduction and Guest Welcome
01:18 Ryan's Background and Journey with Rails
04:00 Technical Insights: Ruby, Rails, and Redis
09:37 Scaling Challenges and Solutions
12:45 Networking and Community Impact
16:15 Development Cycles and Marketing Strategies
19:00 AI Integration in Development
22:47 Origins and Evolution of Knowatoa
25:32 Conclusion and Contact Information
Links:
- Knowatoa
Today I am joined by Ryan Castillo. He is one of the founders of Knowatoa, a platform dedicated to optimizing business visibility and accuracy across AI driven search engines. At Knowatoa, he leverages his extensive experience in data science and software development to help businesses navigate the challenges posed by AI search technologies, ensuring they remain competitive and well represented online. Welcome, Ryan. Glad to have you on
Ryan Castillo:Yeah. Thanks for having me.
Brendan:We talked about this a little bit ahead of time, but,, I actually got that, that from ChatGPT 'cause I think that . It's a pretty good synergy with what we're gonna talk about today. So is there anything you want to add, I guess to my intro for you?
Ryan Castillo:No. I mean, I think that's right on. I mean, that's kind of, our, like quick five second spiel is like, we're tracking how brands are represented by AI search. And it started, with our, with our other company Wafris um I think we had done a ChatGPT search to see if he knew about Wafris and we got a bunch of like Wa waffle related results. So, our initial prototype, our initial interest in the project kind of started off like that kind of joke is like, how come ChatGPT and all these ais are like misrepresenting not only our businesses, but you know, friends of ours as businesses as well. So.
Brendan:Yeah, it is an interesting problem. And before we maybe dive too much into Knowatoa I've got a kind of the kickoff question that I've been using for, guests, which is around our theme of Rails business. What is it about Rails that attracted you? What made you choose Rails and what's kind of kept you with Rails, you know, for these products?
Ryan Castillo:Oh man, that I'll try and keep the short version. I was actually in the biotech field and I was doing Java and we had like pharmaceuticals, and biotech companies as clients. And, man, this was, oh, well, almost two decades ago. but that was starting to kind of be more of a bore and . At the time, I was attending some user groups and people started to talk about not only rails, but like GitHub and like all these new technologies that were just based on rails. And so, and of course it was built on top of Ruby, which I found to be much more elegant, And not having to wait for compile time and those kind of things. And so it just turned out to be a very enjoyable developer experience. So, I think one day I decided that I didn't really want to stick to the management path that I was on at kind of consultancy that I was on at the time, and wanted to double down on rails and I've been kind of running with it ever since.
Brendan:Nice. That's awesome to hear. you guys are still using Ruby and Rails for wafer and Noah to, I assume.
Ryan Castillo:Yes, we are. So, and to kind of guess go back more into that background Mike and I, my, Mike, my co-founder and I are both developers by trade. And we actually met through one of those Ruby and Rails meetups like way back in the day. But we've also kind of, I've kind of let my career drive itself from More of a learning perspective than a trying to solely focus on being a developer. So there were times where, I wanted to learn more about business. So I took, like Amy Hoy's 35 by 500 course. There was a time where I wanted to learn more about the marketing and sales side. So I took a job at a an enterprise company. And throughout those experiences,. Just having rails and, the developer background with Ruby just kind of, seemed like a superpower. So it was just a technology that both Mike and I and knew and it was something that we've always been able to return to. So yeah. Even when we were building Wafris there were some other languages that we turned to, but we always kind of base things on Ruby and Rails and I could go more into that if you want. Those are some like, kind of technical details, but, I'm always happy to talk about it.
Brendan:Yeah, I'm definitely not shy about getting into, more technical details. I mean, one thought for me right there when you were talking was, what are some things that worked well about Ruby and Rails and what are some things maybe that didn't, that maybe pushed you towards other technologies for either one of the products?
Ryan Castillo:Right. So, the big thrust for Wafris was, and just to fa familiarize everyone with Wafris, because I know we talked about Noah to initially wafer is our firewall product. It's kind of like what we started the company on. And the thing there is what we saw was most firewalls were being offered at the enterprise level. Even with CloudFlare, you will, you'll start off with a they'll give you the firewall and stuff like that. But as you get to the higher tiers, it just gets more and more expensive. So, we saw an opportunity to have something like Wafris that, you know, fit between CloudFlare and something like Rack Attack. But what we wanted to do is we kind of wanted to be everywhere at the same time. Not only serving . The Rails community, but also the Laravel community and other, some of the other like technology stacks. And so we came up with this idea of, hey, if we can come up with a firewall client like that can sit anywhere on the stack, we can spin out these clients very quickly. And so our initial premise was like, let's base this all on Redis. Let's store the rules on Redis. Let's store the data in Redis and those kind of things. And. Also store the firewall logic in Redis. And to do that Redis kind of has this plugin architecture where you can write everything in Lua. And so we wrote a giant Lua script that got imported into Redis and processed all our rules. But the way we actually developed that is since we were both Ruby and Rails developers and had little experience with with Lua, is we built everything in Ruby. Then use, ChatGPT to kind of translate it along the way. And, we hit some rough edges here and there, but eventually we were able to come up with a pretty decent Lewis script to, to handle all of that. So it, again, for us, I think to answer your original question, going back to Ruby has just always been like an elegant experience. It's very readable language, if Mike builds something it's something that I could easily go over, months later and figure out exactly what he intended. And the same for him and with the stuff that I build as well.
Brendan:Yeah. And one thing you mentioned there about having that script in lieu of, but you started in Ruby, like I, I don't know how many times I've heard other developers talk about that same process of write it in one language but have ChatGPT, convert it to the other language that you need it in or whatever.
Ryan Castillo:Yeah. And that's one of the nice things about Ruby is like you can just do pseudo coded, like, and it's so easily and now of course with like Cursor and all the other AI tools all you have to do is like kind of write the comments and instruct the AI and they'll generate the code for you. But at the end of the day, you kind of have to understand what's going on and for us, ruby has always just been very easy to read.
Brendan:Yeah. Yeah, definitely. And maybe that's a good point or place for us to, transition into Knowatoa a little bit since there that is more AI and search focused. And we've been talking a little bit about chat, GPT and AI tools. I guess one question I have about that specifically would be like, how does, how are you using rails with that tool and AI and such?
Ryan Castillo:Yeah. For that a lot of that comes down to APIs. So, with noa, toa, all these . All the AI tooling have their own APIs, so we're able to kind of interrogate the prompts and stuff like that and return those results back. So that's just standard Ruby and Rails like integration with APIs there, there's nothing like fancy that we're doing there. Yes, there's like a lot of data and there's a lot of processing and those kind of things. So we have to kind of push that, that stuff out to Sidekiq jobs. And we are starting to kind of push the edges as . As to the amount of data we're storing, but it's nothing that we've found that Ruby and Rails can't handle. One thing that has kind of been beneficial for us is because Ruby and Rails has been so developed we've also known a lot of developers within the community. Over the last, Decade plus. And so, finding help. To kind of work on some of these projects has been pretty easy. And these are pretty senior level developers where they can come in and not only understand what we're trying to do with a product, but and the code and stuff like that, but also kind of advise on like, Hey, this is how we think the UX should work. Or, Hey I worked for this other bootstrap company and we were able to build this kinda reporting and these are the edges that we ran into here. And so. We're able to kind of benefit from, the experience of like senior level developers to, help move the products along.
Brendan:Yeah, that, that's a really big thing, especially, where I'm at. I've got a pretty big team, but not everyone's senior. I've got varying degrees, and we've recently started working with more senior developers, and it is, well, obviously we want to, nurture and grow, junior people. There's a huge gap that you don't always realize until you've got . Big data problems or things like that, that you just, you realize you really need that senior talent. And I would be interested to dive in a little bit on like the database. It doesn't have to be specifics exactly, but like, what are, and this maybe goes into another question I was gonna ask, which is, what are some challenges you've had with the businesses, whether it's rails or not? Sounds like, having a lot of data is definitely one of 'em. Like, what have you found that works? What are you, what are the challenges you've found there?
Ryan Castillo:Yeah. So, the most of the scaling stuff like we really encountered was more on the wafer side. And that's primarily because, we had made an assumption. About Redis. The assumption was that Redis was everywhere, and so when people were to come and get the waf they'd just be able to bring Redis with them. But what we actually found was like a lot of our customers didn't want to be like Redis administrators, and so they didn't want to have to like bring the Redis. They didn't wanna have to pay for two things, those kind of things. And we also just noticed there was more of this. That we were living in a non heroku world. Like we, Mike and I had come up with Heroku and so, if you want Redis just there, but everyone else had been using these services like fly or it's more of like a di distributed environment. So with that, we pivoted technologies and then we decided on like, hey what would this look like if we just took Redis completely out of this? If Redis is the bottleneck technology wise and potentially speed wise what if we just put the rules directly with the application? By doing that we decided to go with SQL Light. And so, with that we kind of built this distributed architecture where no matter how many instances of your application got launched, it would always bring a SQL L database with it that matched all the rules for your firewall and that kind of thing. So there were some cool scaling things that we did there to kind of speed things up. One of the big gotchas that it really boiled down to was like SQL Light just had to be read only. And so even with SQL Light against Redis and we have a blog post that covers this whole thing I think we were, yeah, I think we were able to process like, several thousands of magnitude faster worth of requests with SQL light than Redis.
Brendan:Yeah. That's pretty cool. I remember when that blog post came out and it kind of felt like it also was aligning with I'm forgetting the developer's name who's can been kind of big with SQL Lights and it feels like SQL Lights' kinda having a renaissance a little bit too.
Ryan Castillo:yeah, you're talking about Steven.
Brendan:Yes.
Ryan Castillo:I can't remember Steven's last name, but yeah, he actually, Steven actually recently released a sequel light course with,
Brendan:that. Yep.
Ryan Castillo:with one of our friends, Aaron Francis.
Brendan:Francis.
Ryan Castillo:Yep. So I haven't had the chance to check that out, but I think I had talked with Steven at Rails World and so a lot of the stuff we were doing, was kind of some of the stuff that he had been advising on. And I think Steven and Mike have talked a number of times as well, so.
Brendan:Yeah, that, that's awesome to hear. And I think it kinda just gets back to what you mentioned too about like all that senior talent you've had access to. It's. I don't know if Steven exactly. Aligned with people you knew before, but , I found this too. It's really, you know, networking and just like getting your name out there, but also like, keeping in touch with people. It's such a important thing for probably, for bootstrapping especially, right?
Ryan Castillo:Right.
Brendan:you don't have a ton of money to, to hire all these people, but like to get the help is gotta be really great.
Ryan Castillo:Yes. And that, I mean, this is, something that I've kind of always been an advocate for is just like kind of nurturing your network and nurturing like the people in the community. We, I used to help run the, one of the Ruby meetups locally here in, in Virginia Beach. But, I guess just nurturing those contexts and just trying to be helpful, always creates opportunities later on. Opportunities for collaboration and opportunities for your career. I think in the, in like the prerecording, we talked about how we had met at Madison Ruby, like a decade ago. And at the time I was actually kind of hustling as a consultant. I was trying to fi figure out like, Hey, how can I, scale up, making this money consistently. And what actually came, yeah, what actually came outta that conference was working with the test double guys.
Brendan:Yeah.
Ryan Castillo:I had a previous relationship or, previous engagements with . Justin Searls. And because that went so positively, they were able to bring me on as a sub contractor for several big gigs. And that kind of, propelled my career forward for several years. So, and that was all just born out of like, conference connections and those kind of things. And so, like you said about knowing senior talent, I think it's just a cosm of like to know people in the community and trying to be helpful to them, has always allowed us to, create these opportunities for their theirs. And, my own projects
Brendan:Yeah. Yeah, definitely. Yeah. And you'd mentioned earlier too, 30 by 500 and I think that might've, if I remember right, that's kind of what you and I connected initially, is we were talking and then we both found out we had been doing that and, that was like you
Ryan Castillo:Yeah.
Brendan:jumping off point too. And it's funny because there's so many other people now over the years, again, that was more than a decade ago that you and I originally met. And I've met so many other people that have done 30 by 500 just in the Ruby community. That's been kind of interesting 'cause that's like a whole nother networking effect or just community I guess.
Ryan Castillo:Right. And I think, I think the Laravel community is kind of seeing this now. But I just remember, and then, with the Ruby community, I remember it being very Kind of indie, like, hey, we're not, this isn't necessarily a corporate technology. Like I mentioned, I had done Java before, and so it kind of seemed like, hey, all the guys that were like developing and running their own consultancies, or they were doing freelance or, working remote, everyone in those circles were doing, was doing Ruby and Rails and stuff like that. And and yeah, the same thing. In fact I'm not sure, but I'm. Maybe Amy Hoy, who was the creator of 30 by 500, maybe her, like some of her products were built on Ruby and Rails.
Brendan:I remember right. Yeah, I think
Ryan Castillo:I might be misrepresenting that, but yeah, so I think there, there was just this kind of vibe and energy around the technology at the time that was appealing.
Brendan:Yeah. Yeah, definitely. A follow up question I have would be, I guess, related to that in some ways, you've already talked about . People that have helped you out, but how much development work are you and Mike doing? I assume quite a bit still, right? It's not, you don't have a huge team or anything behind you?
Ryan Castillo:No we, so we we split things into cycles.
Brendan:I
Ryan Castillo:and I think you and Ryan were talking about this in, a previous episode
Brendan:up basically
Ryan Castillo:yeah. Shape up. And we've kind of experimented with that. But one big thing that, we've had to focus on and we kind of use to our advantage is just focusing more on the sales and marketing side. So we kind of try to . Do that very consistently and then release, features, in the off cycle. So we're just coming off a cycle now where we're just like, . We doubled down on marketing on LinkedIn connecting with our users, trying to get them on meetings and those kind of things. And now we're entering in a cycle where we're doing development and feature releases and stuff like that. And that's just the kind of the nature of like, running a SaaS company, especially a small one. Where Yeah, it'd be nice to like kind of code all day and like release these features, but if we don't have the users and the actual people paying for the product, then you know, we're not gonna survive, so.
Brendan:Right. Yeah, that, that's an interesting take on it. Obviously, like you said, you've gotta do the marketing 'cause where the user's gonna come from otherwise, But just having that framework of, correct me if I'm wrong, but it sounded like you're saying you basically do a six week cycle, or maybe it's not six weeks, but you're doing a shape up type cycle that's marketing focused and then you do a development one. Is that kind of how you're doing it?
Ryan Castillo:Not, it's not completely written in stone. We're still kind of experimenting with it. But we're trying, so I think prior to this, we were trying to really get no, a toe out the door. So all of our cycles were more development. But the beginning of this year there was a focus more on The marketing and sales side. Not to say that we weren't doing any development because there we were, I was writing code for more of like our back backroom systems, like tracking users, enriching user information, that kind of step kind of stuff. Mike actually launched a a like a. An SEO pay tracker on LinkedIn. But he built a lot of that on Ruby and Rails just so, he could like recycle that report at any given time. And so that, that's actually, going back to like why we still use Ruby and Rails. It's like we're able to kinda like scale up some of these smaller projects really fast. Us.
Brendan:Yeah.
Ryan Castillo:As an aside to like our bigger efforts with like wafers and Noa.
Brendan:Yeah, totally. Yeah. And another question I guess with development related, we mentioned cursor a little bit earlier. I assume you're using tools like that, but how much has AI helped you? Obviously Ruby and Rails is good at doing things like we've talked about, but now we've got AI in the mix, which wasn't there when you and I first met.
Ryan Castillo:Right. Right.
Brendan:And so now, I've been messing around with a little bit. It's good. In some ways it's not good in other ways. Like, how has it helped you?
Ryan Castillo:It, I mean, quite frankly it's kind of been a core in our business. Mike was Mike was much more of an early adopter than I am, so I trailed behind in the technologies that he uses. But because we utilize like a lot of the AI tooling, we're able to, it kind of move a little bit faster. So it's involved like, like every step of our process from, doing research online to being in, actually integrated into our products to generating images for, our newsletter and stuff like that to building specs for some of our code. It's just . Heavily involved. This is one of those cheats that we've, kind of taken advantage of to kind of bootstrap these two products. And yeah, so it, it's just, I mean, I. I think I use, I actually, I don't think I know a hundred percent, but I'm, I don't even think I use Google anymore. I just, I'm always in either chat, GBT, or cursor or, my primary editor is Vim, but I even have like, copilot integrated into that. So, yeah.
Brendan:Yeah, that's funny. This is a little bit of a side note, but I'm a Vim user as well, and I haven't integrated copilot, but I've been told about super Maven, which is I think more like cursor, but I've actually never used copilot, so I obviously that's cursor too, but not exactly.
Ryan Castillo:I mean it's um, so this my workflow . Typically tends to be like, I'll be in Vim and I'll, if I type out comments and stuff like that . Copilot will help generate some of the code. And I think I'm using chat GPS integration for that as well. But if it gets too complicated, like where I feel like I need a pair. Then I'll actually go to Cursor. I'll just, go back to the terminal and I'll just like cursor dot, and then I'll open up the entire project. And there the kind of like the interaction with the AI begins is like, Hey, I'm trying to write outta the spec for this, can you generate that for me? Or, Hey, this code looks gnarly. You know, Is there something to make it look a little bit more elegant or. Or maybe I'm hitting a roadblock or something like that and I just need to rubber duck something. So I start chatting with Cursor and it's like helping me trying to generate all that code and stuff like that. it's definitely been immensely helpful for a number of reasons.
Brendan:That, that's cool to hear because I've essentially been doing the same thing and I was trying to decide , is it because I'm a Vim user or what, but yeah I'll do most of my coding in Vim, but yeah, when I feel like I need to have a conversation, I moved a cursor just like you just described. And hammer things out. And then once I'm at a good point, I go back to Vim and kinda keep cruising. So that's kind of cool.
Ryan Castillo:I mean I know there, there are some more purists where they use like ChatGPT and they'll say like, Hey, I'll let it generate the code for me, but then I'm gonna like kind of type it all in so that I understand exactly what's going on. I, I. I'd love to be like that. I just don't have the time, frankly. So, I think at the level that I'm doing it, I still have like a deep understanding of what's going on with the code and like what we're pushing. And so that, in turn will make things like easier to debug later on.
Brendan:Yeah. Yep. Well, I know we've kind of gone on a little tangent on AI and that's totally fine. It is related to Knowatoa, so . Maybe let's kind of circle back to Noah to a little bit. Is there anything else you want to add about it or like other things that we haven't talked about yet? Around the, that tool?
Ryan Castillo:No I do think, I think on a previous, this might've been the podcast that you guys had or the episode you guys had with Jason, you guys were talking about like, Hey, you know, it's . Been really hard to find like a product idea. Like I think someone mentioned like they've been searching for a product idea and it just hasn't come to them. And I think as developers, I think what tends to happen is we tend to kind of focus around tooling and stuff like that, especially our own tooling. Instead of like the. The business verticals that we're operating in. So what's worked for me and what's worked for us was the part of the reason why Wafris was born is we were at a data security company. And, we were selling, security products. And so we saw the gap in the market there. And so we were able to say like, Hey we know about the security vertical. We know there's a need for firewalls. And so like, this is how we're able to kind of find the intersection between everything we know we can speak to security people, we know we can sell this to developers and that kind of thing. And so. That's how Wafris was born. The way Noah to was born, which I kind of started . I highlighted very briefly was, we kind of just started asking ChatGPT. Like, Hey, what do you know about wa? What do you know about HubSpot? What do you know about geo coio? What do you know about Judoscale, Adam's company? What do you know about all these different things? And it just was not returning like in informative results. And then so we started showing these to our friends, and then they were like, oh yeah, this is a problem. And just kind of observing that and kind of continuing to kind of build on that report. And that, that pain that, our friends were encountering with their products and stuff like that. And just seeing the momentum build from that is how we came to, to double down on Knowatoa. And so it's, I think I. I guess one thing that I would advise, the listener to is like, you don't necessarily kind of go out for looking for a product. You kind of, and this goes back to 30 by 500, right? Is you like, you know that your community, whether that's the developer community, whether that's for us, the marketing community, whether that's the, security com community and you just kinda listen to people and listen to the pains that they're encountering. And based on that, that's where you can kind of figure out whether an idea is worth like investigating and building something small around.
Brendan:Yeah that's really cool to hear. Thanks for, giving some more detail around the origin of both products and it feels like maybe that's a pretty good . Place to wrap it up.
Ryan Castillo:Cool.
Brendan:Thanks for being on Ryan. It's been great talking to you. Is there anywhere that you want people to find you on the internet?
Ryan Castillo:Yeah. You can, right now find me at or contact me atRyan@noahtoa.com. I'm on Bluesky, I'm on LinkedIn. I'm actually more LinkedIn more active on LinkedIn nowadays just because that's where we're finding most of our users. Yeah. But if you reach out to that email or find me at RM still on Bluesky, I am happy to respond there as well. And yeah, just thanks for having me, Brendan. Really enjoyed it.
Brendan:Thanks, Ryan.
People on this episode
Podcasts we love
Check out these other fine podcasts recommended by us, not an algorithm.
Code with Jason
Jason Swett
Standing in the Fire
Very Good Software
Remote Ruby
Chris Oliver, Andrew Mason, David Hill
IndieRails
Jess Brown & Jeremy Smith
Startups For the Rest of Us
Rob Walling
Vibing with AI Code
Brendan Buckingham