In this MonkCast episode, Martin Rojas, UI Architect at PlayOn Sports and ReactATL Organizer, joins Kate Holterhoff to chat about the tech meetup scene, leveraging AI to enhance sports broadcasting, AI Guilds (not Dungeons & Dragons!), Micro frontends, React, and More!
Links
Transcript
Kate Holterhoff (00:12)
Hello and welcome to this Redmonk conversation. My name is Kate Holterhoff, Senior Analyst at Redmonk, and with me today I have Martin Rojas, UI Architect at PlayOnSports and React ATL Organizer. Martin, thanks so much for joining me on the MonkCast.
Martin Rojas (00:26)
Thank you Kate for having me. It’s always a pleasure.
Kate Holterhoff (00:46)
Yeah, super excited to talk to Martin. So Martin and I both live in Atlanta and we run into each other not only at the sort of local meetup scene, but also at a lot of the big conferences that focus on the front end community. So I’m to start with your role as part of the React Atlanta group. Can you talk about what that is and what purpose it serves in the community?
Martin Rojas (00:46)
Of course, actually was very interesting that kind of was thrust upon me at the beginning of the pandemic. Naya, the previous organizer, had to leave Atlanta for some serious reasons. So I was kind of thrust into doing this work after only volunteering a couple of times.
right as the pandemic started. So had to kind of figure it out on the fly. We ended up doing a lot of Zoom meetups and that kind of stuff. I was able to keep it going through a pandemic. And as I did that, I kind of saw the value of a lot of people that were shifting careers, that were kind of trying to find that community. And that’s, to me, that’s what the biggest thing is, right? There’s a lot of people that will reach out when they’re in the middle of a job transition or they are…
you know, they’re trying to figure out what their next steps are, right? Where they should be working towards, or there maybe they’ve been stuck at half mid-level of their career and they don’t know how they can get to that next step where they should be learning. So from that, it’s been really kind of rewarding and interesting in hearing a lot of different perspectives where people have been coming from, the kind of work that they do and what they’re encountering.
Kate Holterhoff (02:07)
And so you were involved pre-pandemic and post-pandemic or did you just get involved more recently?
Martin Rojas (02:07)
It was actually I had reached out a few months before the pandemic and had done my first talk because I was like, man, I’m a senior. I want to kind of share back to the community. You know, I never came out. So I had done one talk that I really enjoyed it. So I had reached out to the organizer at that time and I’m like, hey, what can I do to help?
How can I be more involved? And I helped in a couple more meetups. And then she’s like, hey, I need to go back to Canada, go back home. And you want to take it over?
Kate Holterhoff (02:48)
And you said yes?
Martin Rojas (02:50)
I’m like, yeah, sure. I’ll give it my best shot.
Kate Holterhoff (02:56)
I love that. Okay, so you were all in. And so you’ve seen it before the pandemic and after the pandemic. I’m interested in what you’re seeing there because I know the meetup scene, the conference scene really took a hit during the pandemic. And in Europe, it’s back up. I’m hearing sort of mixed accounts. about the US conference scene. What are you seeing around meetups? Do you see that attendance is back where it was pre-pandemic? Have things changed? what’s a pulse check?
Martin Rojas (03:23)
It’s definitely coming back. I think people have been coming back this year a lot more. It’s a very different kind of vibe where people who are coming out right in you know in a good way or bad way pre-pandemic or even during the pandemic it was a lot of a lot more of people that were ⁓
hey, I’m just starting or I’m switching careers or I’m switching this. I think in the last probably six months, we are seeing a lot more of people that are already working start coming out because there’s a lot of things kind of changing. a lot of them are, same thing, a lot of people come out when they’ve been laid off and are looking for jobs and are looking for in there, but it’s not always that.
Kate Holterhoff (04:12)
Yeah.
Martin Rojas (04:15)
which that was the only time before where you would see people mid-career or seniors really come out. Now I think that when you’re posting topics about some of the new technologies, things are going so quickly that people are feeling like there’s a lot they need to keep up with or they need to do to not get left behind, right? AI, of course, being the big key term there, being in there and knowing how they can use it to understand that there’s so much to…of drink from a firehouse on that.
Kate Holterhoff (04:49)
that’s so interesting. So your take is that there is more interest in the meetup scene because things are moving so quickly around AI. Do you think it’s tied to, you know, we’ve heard a of rumblings that folks are losing their that the are frozen. I tend to blame that more on the post-ZIRP era that we’re in. You know, I hear a lot about offshore, so I’m not ready to point my finger at AI. However, maybe you are. Like, what is your sense? Do you see the relationship of hiring around developers to be reflected in attendance at the meetup that you run?
Martin Rojas (05:28)
In a way and because of before you know even during the pandemic or before after If you had a developer you already had a career you were pretty safe in knowing that hey If I get laid off or something in a month, I’ll probably have a job may not be the best job, but I can pick up something, right? I think everybody’s been seeing that, it may take six months to a year to find that career. So everybody’s being like, okay, I need to start laying groundwork or I need to keep my options open because I don’t know what’s gonna happen and I can’t like start from zero. So it’s driving a lot of people to reach out and to start building those networks.
Kate Holterhoff (06:15)
that makes sense. All right, and I guess before we move on to talking about some of the more technical subjects that I’m hoping we’ll get to here, do you have any advice for other folks who are in the meetup space that run a meetup and that are or maybe thinking about becoming involved in one?
Martin Rojas (06:15)
I mean one of the biggest things if you’re looking to get involved a Lot of the people that are running the meetups are because they just kept showing up It’s It is it is a lot of work. It is you know, it is It’s in there. It’s a lot of different skills and It came out to us that us being able to share that load makes it so much easier and less of a burnout
where we can have consistent, we’ve now been able to do almost two years of a monthly meetup. And there’s been, yeah, at least monthly, but every other month, but we don’t have to all plan, we don’t have to all be looking for people. Sometimes I’ve had some, sometimes where my personal life has kind of caught up and there’s been like two months where I’ve been able to just step back and it still keeps going, right? It’s still, so finding you know finding those cross-reasons we haven’t given up the identity of our meetups We are just like and we always tell people hey, we’re working together. Just being honest You know we’re working together if you follow one of us you’re following all of us so Everybody kind of gets found out you know gets notifying you get a lot more people out
Kate Holterhoff (07:51)
Got it, got it, got it. OK, that helps me. All right, so let’s talk a little bit about what you’re doing at Play On Sports, because what is very interesting to me about your role is that you are working as a UI architect. Because I spent a lot of my time thinking about what it is that designers, that UX engineers, and the front end are seeing in our current moment. How is that affecting how we use CMSs? How is that affecting just the way that you’re using Vibe Coding. talk to me about what you’re doing at PlayOn Sports that is teasing out this nascent relationship between AI and the UI; the front end.
Martin Rojas (08:21)
You know, we’ve adopted, we have a new CTO and he’s adopted a, he’s really a big proponent of the guild framework for, you know, for a group. we’re building out a lot of guilds. One of the main ones, and it’s really big, it’d be a big push, it’s an AI guild. But one of the more interesting things is that it’s not just a guild inside of the technology organizations. It’s something that is being open. You know, we’re…
sort of creating it as a way to share the knowledge that we’ve been building up as we’ve been learning about AI inside a technology group to our entire company, right? We’re launching it to be a point so that we are creating, you know, one of a person in our AI guild, is the head of learning for the entire company. So our hope is that as she’s doing trainings in a lot of different parts of it, she does leadership training, does that AI alert trainings will become
will start becoming part of it so that they’ll be available to everybody in the company, right? Also, marketing, sales, they’re all looking at AI tools to help their work go faster or work better, but they may not have the experience of knowing what questions they should be asking vendors or knowing if they’re missing something. So the way we’re building the guild is that,
We’ll have our experts, we’ll have our point. We have legal as part of it, we have security. And they know that they just have one, I’ll be the main point of contact, but they can reach out to me or to the guild. We have a Slack channel being like, hey, I’m trying to use this tool. Can you help me be that supporter, help me make sure I’m asking the right questions, that I’m not gonna get caught up in there. We’ve created sort of an evaluation framework of like,
when a business director or manager is looking at in there, sort of kind of a checklist that they can go down and be like, this is what I was looking for. This is what I should be aiming to do in this POC stage or my first rollout or for rollout wide to make sure that we’re not trying to build something that is not actually going to do what we want it to do, waste money. Cause that’s something with AI, right? It can looks really nice, looks really cool, but it may not actually move the needle on what we’re trying to do as a company as a whole right in the company objectives.
Kate Holterhoff (11:02)
and for folks who aren’t familiar with the guild system, it sounds very Dungeons and Dragons. Can you define that for us?
Martin Rojas (11:02)
The idea of a company guild is sort of creating focus groups. So whether they are management guilds or in this case, an AI guild, it’s sort of to create a framework where you can have your own meetings. You can create learnings. People can bring it up. It’s a way of kind of showing like, hey, if you’re interested in this topic or in this,
this is the space where we are going to help you expand that knowledge in there. And of course, can, depending on the amount of effort you’re putting in there, it can be a lot more than that, right? can be a management guild. can be something as like, Hey, in this guild, we’re going to be doing a book club on like management, you know, a management book and we’ll kind of describe it. Or if somebody has learned something and they want to do a presentation, it creates kind of that framework for that and in there.
so they can be as defined as, you know, as structure as we’re going with the AI guild, but it gives you something that you can attach kind of those responsibilities to. You can have more of a core group, thinking almost like a employee group, it’s loose, but the idea is that it provides a space where you can, you know, have documentation, get resources, and… allow employees if they’re interested in learning something to know how they can look at where those resources exist within a company.
Kate Holterhoff (12:31)
I love that idea. OK, so that makes a lot of sense to me. So it’s like a continuous learning. gives some community, provides documentation so that folks who maybe aren’t a member can then look over and say, hey, what are they doing over there? That’s great. Yeah, OK. So there’s an AI guild. You’re learning a lot. And you’re leading this guild. Is that right?
Martin Rojas (12:38)
Yes, I’ve been given the opportunity and given the time in my work to make this be about half my, 50 % of my work, which at first I was like, oh wow, this is, I was like, is it really gonna take this time? But as we more roll that out, be in there, because it’s given me the time to reach out to different parts of the company. One of them when I was working with our
internal communication person, and we’re trying to figure out how to publicize it, how to roll it out. She, one of her comments from like a non-technical perspective is like, well, if I see something AI, I’m going to think that has nothing to do with me, right? That’s a product, that’s engineering that has nothing to do with me. So it was very interesting, very different perspective. So in same thing, and being able to reach out outside of the engineering and product.
organization, it’s helping me shape how I do the message and allow it and tell, know, and making sure the message is clear that you don’t have to have any knowledge in there. As long as you’re interested, you’re welcome and we will help you learn and become like, we have these, I created this term called AI champions, right? So any department, if you’re interested, you want to be a point of contact for me.
to help to know like, hey, you’re really interested in AI and there’s somebody that if I find some good resources or a good online conference or something like that, I can send to you and you are more receptive to that. So it creates that. So it’s like, you’re not taking fully all of your time, but you’re a little bit more interesting. It’s helping me create that kind of link and flow back and forth. That’s so exciting. OK. Yeah, I think we hear a lot about the CEO who just late
Kate Holterhoff (14:30)
That’s so exciting. OK. Yeah, I think we hear a lot about the CEO who just like late at night send a Slack message to someone in engineering and being like, hey, can’t we just automate this role? you know, these are all these stereotypes about folks who are sort of undereducated about AI, treating it as this magic wand that’s going to solve all their woes, to continue the Dungeons and Dragons metaphor. And it sounds like what you’re doing, though, is totally different from that, where it’s like the leadership has said, hey, we’re going to make this an organization-wide initiative in order to incorporate AI in a thoughtful way that is bringing in a lot of stakeholders and making sure that we’re doing this in a way that makes the most sense for our particular use cases. And I love that. That’s great.
Martin Rojas (15:04)
Right. And it is. it’s actually AI is such a good fit for what we PlayOn Sports we do in there. you know, context, we work with high school sports nationwide, right? We work with the athletic departments. We work
We we work with students, we do ticketing, streaming, and all that. And as sports fans, kind of what we want to do is we want to be able to bring as much of the stuff you see on like an MLB broadcast or that you see on a college level broadcast that has a lot of work onto it. We want to be able to offer that at high school level. So when we do a live stream, we’ll be streaming every Friday night.
for football, for Friday Night Lights, we’ll be streaming over 5,000 live streams. There’s no way you can do the manual parts that you see in a broadcast in there. So AI is a perfect use case where it’s not really taking everyone’s job and things like that. if we can figure out the right way to frame it and do it, we can provide so much more for the high schools and the parents and everybody that’s watching, we can give them so much more.
Kate Holterhoff (16:32)
And can you talk a little bit about specifics there? Like, how are you leveraging AI as, part of your operations?
Martin Rojas (16:32)
Yes, some of the things and some of the goals in there, we’re trying to find a good balance on if you’re watching the games, you’re interested in following your grandparent, you’re you’re following your kid, you wanna see, our goal is to be able to create highlights and be like, hey, my son was a player, I wanna see his highlights for the game.
I wanna be able to create highlights for the game. Wanna create those recaps, those highlights. Same thing nationwide, a lot of people, what were those really good plays that happened over the weekend? you see ESPN has those like 10 top highlights. We’re doing that, it’s a little more manual when they’re like, if we can leverage AI to find those great moments, like so many.
awesome opportunities and things that happens throughout the, you know, throughout a Friday night and so many nationwide. If we can find those awesome things and bring them up and, highlight them, it helps everybody, helps the player, helps the community that’s finding them. So that’s one of the things that we’re kind of working and trying to find it. Also in, same thing, creating
great articles of post games like post game recaps. So it’s a lot of the content you expect and you see after a game’s happening, but because of the volume and coaches are honestly under so much things going on in there and even getting players looked up, we’ve been playing and being able to, you have a roster printed out, being able to just take a picture and have all that data inputted into creating the rosters, right?
Kate Holterhoff (17:56)
Okay.
I see. Okay, that makes sense. All right.
Martin Rojas (18:23)
So yeah, it’s a lot of automations on a lot of things that just helps the coaches just manage an entire thing
Kate Holterhoff (18:32)
Got it, got it, got it. So would they upload, I’m trying to think of, this is still little hand wavy for me. So you’re uploading the roster and then you also, would it be like the transcript of the entire game and then it would pick the highlights out from there? I’m trying to understand how it would pick the highlights specifically.
Martin Rojas (18:49)
The roster and stuff, that happens more at the beginning of the season, right? When you have your team and things set up, but like same thing, you’re doing this as school starting up, you have so many, you’re a coach or athletic director, you may have like five or 10 teams and you’re trying to set up all that correctly. You have a printed list out that you may have made in Excel or something, or you have written out by hand. OCR is so good at being able to take that handwriting and turn it into text and…
Kate Holterhoff (18:52)
Okay.
huh.
Martin Rojas (19:17)
Then you get a first pass and then you can look at it and essentially helps you get 90 % of the way there, right? With AI, gets you 90 % of the way and then that way is just cleaning up, making sure the player spells this card correctly. But it helps that initial setup of the team as the year’s starting so much easier for the coaches, right? Those are some things that AI can really help out with. Okay, all right, yeah. I one of the youth cases that I…
Kate Holterhoff (19:42)
OK. All right. Yeah. I know one of the use cases that I love is, frankly, around this podcast where it take the transcript and then summarize it into the abstract. mean, that is amazing. Because that, for me, was a little bottleneck. I said what I said. I don’t feel like going over this again. I’m thinking that maybe there’s something similar if you’re dealing with broadcast. But it sounds like that’s not quite the use case that you have right now.
Martin Rojas (20:08)
It’s one of the problems we’re figuring out to solve of how can we take that video feed and create that summary.
Kate Holterhoff (20:16)
Yeah. Yeah, yeah, yeah.
Martin Rojas (20:17)
Yeah. Right? same thing. If in MLB, in the professional games, they have somebody that is creating that like transcript of the game by hand and doing that, it would be impossible to do that for every high school on the time. So that’s a perfect use case for AI. All right. So things coming down the road in that case.
Kate Holterhoff (20:36)
yeah, yeah, I can’t imagine. Okay. All right, so things coming down the road in that case. And you’re not using the AI models that can manipulate video. That’s not part of it.
Martin Rojas (20:50)
It’s one of the options we’re looking at. So yeah, I mean, we’re looking at LLMs. The only one that can do video is Gemini at the moment. So that is the only true multimodal LLM that’s happened that’s right now. So it is one of them same thing. Because of the scale and the volume, it’s in trying to find what AI is good at which part of it.
Kate Holterhoff (20:52)
okay. yeah.
Okay, great.
Yes.
Martin Rojas (21:19)
you don’t wanna just throw three hours of video to it and then like at a time, it’s like what kind of pre-processing we can do, what kind of cleanup and to make sure we get that like, that really good area. So it’s, yeah, it’s some of the things we’re definitely evaluating a whole bunch of different processes and ideas and things honestly, all the way from getting just the scoreboard, you know, getting the score accurately, right, for the game.
Kate Holterhoff (21:36)
Yeah.
Martin Rojas (21:47)
We have so many different ways that people are connected to us or sending us information. Sometimes the scores don’t line up quite right. And one of the things we can do is like from the video, can we do that OCR on the video and get the score out?
Kate Holterhoff (22:03)
Got it, got it. At many companies, their first AI project is a chatbot. Do you have a chatbot?
Martin Rojas (22:16)
some of our customer ticketing support, have some AI has been implemented into that. So same thing, are, one of the things that we’re really focusing on is what makes sense to build versus buy. That’s one of the making sure we’re getting good ROI in everything we do, right? Do we get a benefit of buying it versus building it? And where is that business advantage, right? Where do we put our resources?
Kate Holterhoff (22:46)
Yeah, that’s huge. And are you using one model above all others? Are you an OpenAI shop, a Claude shop?
Martin Rojas (22:52)
Not at the moment. We are open and it’s, think the, guidance has been, it’s like, we want to find, we want to use what’s the best at the moment to solve our problems. so yeah, we’re definitely like experimenting, looking at it. And if this is the best model for this week, you know, and next week, another one comes along. One of the kind of guidance that I’ve been, I’m trying to, to push out is that, We need to build with the eye that the model that you’re using should be able to change every week if needed. And everything else that you built around it should not be relying on building things specific for a model.
Kate Holterhoff (23:27)
Okay.
Okay, and last question, then we can move on. But are you leaning towards a serverless solution, getting your GPUs from the cloud, or are you leaning towards self-hosting using maybe even an open source model?
Martin Rojas (23:44)
We go back and forth. It’s too early to say because, and same thing, it depends on the problem, right? Sure, yeah. The things we’re doing with video and things like that, some things we may end up doing locally may make sense to get some hardware to do it, but a lot of other things are on demand and you kind of get that cloud.
Kate Holterhoff (23:51)
Is it too early to say? Okay.
Sure, yeah.
Martin Rojas (24:17)
We’re a cloud shop, right? So that’s our kind of first solution for things. But if there is definitely an advantage, we’re not ruling out anything.
Kate Holterhoff (24:27)
Okay, okay. so Martin does a lot of amazing articles on LinkedIn. And so, I know you’ve written a little bit about Microsoft NLWeb, you’ve written about Google’s Gemini chatbot. So I get the sense that you are testing out different solutions in this space. And even reading some of the academic articles, like some of the Apple researchers. So I will be very interested to hear how this story develops and what you alight on and what use cases you discover because, well, I mean, I’ll say that I’ve been having a lot of conversations around the AI code review tools. So that is particularly interesting. especially for the developers at Play On Sports, that there’s these use cases that are going to be valuable to them and maybe they’re going to want their own ChatGPT subscriptions or Claude subscriptions to use as part of maybe their IDEs. Whereas yeah for the chatbot maybe you’d want a completely different solution on-prem or elsewhere.
Martin Rojas (25:12)
No, and that’s exactly right. We have been using some of those tools we’re evaluating. even looking at even the developer tools, right? Which one is in there? We’re in the process of evaluating them and trying to figure out which works. And one of the things we said from the beginning, it may be that one tool may work really well for our front end developers. Another tool may work a lot better for our cloud engineers or back end developers. So.
Kate Holterhoff (25:31)
Yeah.
Yeah.
Martin Rojas (25:49)
It’s the thing, it’s finding the right tool for the right position and not trying to shoehorn something that works, trying to make it work badly for everybody, right? Yeah, it never works. The developers are going to use the tools that make sense to them. You can try to talk to them from a top-down positioning, but at the end of the day, they’re going to be only leveraging the things that
Kate Holterhoff (25:59)
Yeah, it never works. The developers are going to use the tools that make sense to them. You can try to talk to them from a top-down positioning, but at the end of the day, they’re going to be only leveraging the things that actually work
Martin Rojas (26:11)
Right, so the goal that we have is that it’s in kind of like being open with developers and being like, if you’re looking at a tool, you need to make sure that what you’re doing is not sharing company information, you’re taking care of PII. So kind of instructing them on the things that really end up mattering and that would be a problem if they didn’t get out to make sure that they’re aware of that so that if they’re looking at a tool, they can do that evaluation.
Kate Holterhoff (26:37)
All right, so we talked about AI. I want to talk about front end too, because I know that you have some opinions about React 19. You have some on micro front ends. what are you following in the front end space on the top of the stack client side? What is it that concerns you right now? And yeah, give us some of your hot takes on that.
Martin Rojas (26:40)
Yeah, think it’s one of the, know, React 19, really what that brought was all the SSR, you know, server side rendering, and that’s kind of shift back and forth. That’s something that we do have, and we’ve implemented a lot into our things and kind of seen how that
that works in all Next.js as well as in other frameworks because SSR really it’s same thing with everything is finding the when does it make sense to have servers you go back into having to load balancing and having to pay for that that you didn’t have to do when things were done client-side. If you’re doing a client-side or an SPA app you just had to have a S3 bucket and that was all the server headaches that you had. Everything’s in the API.
Kate Holterhoff (27:29)
Mm-hmm.
Martin Rojas (27:48)
Now you’re back to having to have a server, you may get too much traffic, your server gets bottlenecked and all that stuff. But that’s what you need to do. To me, SSRs, if you have SEO requirement, whether it’s Google SEO search or even now with the AIs doing the scraping, they’re not loading up or
your JavaScript. So they’re just looking at what your HTML is. And that even goes back into accessibility. Accessibility now takes a much bigger term. I’ve even started bringing back the term semantic web. I don’t know if you remember it. That was when HTML5 came out and the reason for HTML5. But the whole purpose of semantic web was that
Kate Holterhoff (28:38)
Yes! wow.
Martin Rojas (28:45)
needed to make your HTML code make sense from just the tags or the way that you organize it and laid it out and it turns out that LLMs are now your readers and that means your website needs to make semantic sense in just the HTML you can’t rely on waiting for JavaScript to load to make it make sense or even the visuals they’re not even looking at the visuals that’s expensive to render
Kate Holterhoff (29:07)
that’s exciting. Okay, so PlayOn Sports is a Next.js shop that you’re saying. That’s, when you say we, that’s… okay.
Martin Rojas (29:10)
We have a lot of different things. So we are using Nuxt, we’re using Next, we’re using React Native. So it’s given me a chance to really touch on everything. And same thing, in some of our tools, they’re gotten big enough that we’re getting to the point of thinking about micro-frontends and how do we do that kind of work.
Kate Holterhoff (29:19)
That was a wonderful segue. Let’s hear your take on micro frontends. I mean, that’s a big buzzword these days.
Martin Rojas (29:41)
Yes It’s it’s something we’ve been going in there It’s been going back and forth and I’ve gotten the opportunity to talk to You know people at the Zephyr Cloud, know Luca Mezzalira the person that wrote the book on on in there and My personal takeaway and from seeing other, you know talking to the developers that have implemented other companies is that? Micro front ends is solving the same problem
of microservices, right? You’re wanting to have your teams. It’s a people problem. It’s not a technical problem, right? From a technical perspective is module federation and those, or monoliths are all configuration changes. They’re build changes. They’re not necessarily code changes. And I think that’s where, that’s been one of the biggest things that I’ve been able to share with my leadership and we have taken.
Kate Holterhoff (30:14)
Yeah.
Martin Rojas (30:36)
you know, they’ve taken up, they listen, we’re in a good process of doing that is that this is a technical solution for a people problem, Micro frontends make sense when you have a team, teams that need to ship things at different speeds, right? In the research, and I think it’s when you have with that mindset that that is how you’re going to break up.
your front end, it’s like, hey, this is the work that this team needs to do and this is where the box is, then it becomes a solution that can scale. But if you don’t have your product or you don’t have the sign, buy in and be part of that breakup, it becomes a problem because they don’t wanna do, but I want this team to do something in this part of the page. Well, that’s another module.
So now they’re back to being linked and you can’t deploy one without deploying the other one and you’re back to where you started, which is kind of what tends to happen, right? That’s from the, know, tech leads and other companies that have tried to roll it out and, you know, bringing it back. A lot of that seems to be the problem that they draw the boxes, they break up the code, but then the next feature that comes along will just start linking things back up again.
So they’re back to where we started, they’re in work, because now they’re doing to deploy in one part, it’s not just deploying the big front end, they have to do three different deployments. One of the things that I’ve kind of brought up is that micro front ends doesn’t get rid of the complexity, it just takes it away from the developers and puts it on the managers and the business leaders. So.
if they’re not willing to take on and figure out those like, now dependencies have to be worked out by the scrum leader, the product owner and the UX. So it’s not the developers that are trying to make sure those dependencies are aligned, but they have to be now be done by your product leadership, your tech leadership, your engineering manager now has to solve those problems. So it’s that whole process of shifting left.
Kate Holterhoff (32:54)
I like that. I think you put that really well. And that’s a really helpful description. And so to be clear, at PlayOn Sports, like AI, you are thinking about it. You are considering micro frontends, but you’re not using them currently.
Martin Rojas (32:58)
We are actually, so we are same thing. we took this as a project and we’re like, okay, in order to do this, we need to make sure our automation is up to par because same thing, the second we go, we, you know, we quote unquote flip the switch, it’s, you know, you’re going from one deployment to 20, you know, 10, 20 deployments, or it could be as many teams as you’re willing to bring on.
Kate Holterhoff (33:22)
Yeah.
Martin Rojas (33:32)
So if you don’t have your automation that you can be safe of like, essentially I hit deploy and I don’t, it deploys, it runs all the tests and I am 90 % confident that what I deploy doesn’t crash. It’s good. So we have been doing the cleanup. We have been getting our testing strategy, know, trying to make sure like, okay, now if we’re going to be doing it in different teams, what does that testing look like?
We have been doing all the work and like I said, to turn on, we move into a monorepo. We’re figuring out what parts make sense to break up code. We’ve already had some of the new features that we’re implementing. We’re already implementing them in different kind of folders and they’ve gotten comfortable that they’re technically deploying already. Even though it’s not like module federation, we do have some micro frontends where it’s one team, they’re doing the work entirely separately.
Kate Holterhoff (34:25)
Mm-hmm.
Martin Rojas (34:30)
And when they’re ready to deploy, they can deploy, they’re still deploying the entire application, but they’re deploying, right? So in a way that’s a micro frontend. We’re already kind of behaving. We’ve been doing a lot of feedback of what learnings they’ve been doing as they’ve been doing that work to make sure that as we kind of keep rolling it out and making a stronger shell and all the kind of services that we need, we are ready to do that. To me, like turning on module federation is almost the last step.
It’s that old DevOps things like it can still be a micro frontend work if any team at any point can be like, deploy and the pipeline, the build pipeline just takes it and deploys it. And if you can do 10, 20 deploys a day and everybody’s working independently and can launch it, that’s a micro frontend. And then it becomes trivial to suddenly deploy it with a module federation.
Kate Holterhoff (35:25)
Okay, thank you for clarifying. All right, I love this. It’s been such a great conversation. I’m going to wrap us up here. Before I do, how can folks hear more from you? Are you particularly active on any social channels and are you planning to speak anywhere in 2025?
Martin Rojas (35:26)
I’m in LinkedIn. I think the best plug that I can do is for Meetup if you look at React ATL.
Kate Holterhoff (35:52)
Yes.
Martin Rojas (35:53)
In the personal, I do have a blog where I kind of put my thoughts, just this nextsteps.dev, same thing. It’s next steps because I always found like a gap in everything was a lot of tutorials or white papers. So I’m trying to write blogs where I’m not really covering like brand new stuff, but things like I’m a developer or I may have heard about AI and that’s like the next thing.
Kate Holterhoff (36:11)
Yeah, that space. That’s a really great idea like that. Yeah, you’re totally right. It’s either like hello world or it’s so in the weeds as you’re reading the documentation Yeah, that I think that’s a good place to be, you know devoting some effort into thinking through how to help folks Amazing. Okay. Well, I have really enjoyed speaking with you Martin again My name is Kate Holterhoff senior analyst at RedMonk If you enjoyed this conversation, please like subscribe and review the MonkCast on your podcast platform of choice If you’re watching us on RedMonk’s YouTube channel, please like subscribe and engage with us in the comments.
No Comments