Ep. 31 Building Love at Scale: Inside Tinder’s Cloud Infrastructure

00:12 – Introduction to Chris O’Brien & Tinder

Matt Pacheco
Hello everyone and welcome to the Cloud Currents podcast: the podcast that navigates the ever-evolving landscape of cloud computing and its impact on modern businesses. On this podcast, we typically discuss big trends around the cloud industry. Such as, such as AI, cybersecurity, costs in the cloud, talent shortages, literally everything under the sun related to the cloud.

I'm your host, Matt Pacheco and I am the head of content marketing at TierPoint. And today we're joined by Chris O'Brien, Director of Engineering at Tinder, where he leads both the cloud infrastructure and data engineering teams. With over two decades of experience in technology leadership, Chris has been instrumental in scaling Tinder's infrastructure to support billions of daily interactions while achieving 99.99% reliability. That's a hard one to get.

Today we'll explore the challenges and solutions in managing cloud infrastructure at a massive scale, dive into how Tinder implements AI and machine learning in the cloud, potentially, and discuss the future of cloud architecture in social platforms like Tinder. So, Chris, welcome to Cloud Currents. We're really glad to have you on today.

Chris O'Brien
Yeah, thank you for having me, Matt. I really appreciate it. I'm happy to be here.

Matt Pacheco
We can't wait to hear what you have to say and learn a little bit about you and also what you're doing at Tinder. It's, it's a very interesting app and organization, it's global. So, we're going to ask you a lot of good questions about how to manage some of the complexity and some of the interesting innovation happening in the space. So really appreciate you.

Chris O'Brien
Yeah, sounds great.

Matt Pacheco
So can we start with your personal career journey? We'd love to understand where you started and how you ended up at Tinder.

Chris O'Brien
Yeah, for sure. So I would say I've taken a little bit of a winding road to my current position. So my first job out of high school is actually as a firefighter/EMT. So I did that for seven years. But I would say that compiling the Linux kernel from scratch in my parents basement throughout high school ultimately kind of led me to some technology opportunities with some people I had worked with in the past. So I initially started at a website called FileFront. So back then in the early 2000s, computer gaming, specifically like downloading mods for games or skin packs, was extremely popular. And FileFront, the name of the website, was one of the most popular websites for that, for you to basically upload those mods and skin packs and then download them. And I actually initially started as like a gaming news editor.

I, I don't know if you were really into PC gaming, but it was hard. Okay, so back then like you have, I would have to go to like websites like Voodoo Extreme and IGN to like get like the latest gaming news. Well, on FileFront we had a, you know, a gaming news web, you know, like section. And so I was the one posting all the content. And then, you know, my friend, mentor/co-founder of this company approached me and wanted me to take over as a CTO. So I took advantage of those Linux skills and ultimately took over as CTO of this very small file sharing website. And you know, here I am today.

Matt Pacheco
Awesome. So tell us a little bit about your introduction to Tinder as well.

Chris O'Brien
Yeah, yeah, I, I knew of Tinder. I had not used Tinder, but I worked very closely with, I had met someone when I first moved to LA at a startup. He was a backend engineer that had joined from Amazon and we kind of, more or less, in tandem at this company called Super League Gaming, had run a kind of the back end and infrastructure for this kind of startup which eventually went IPO and we met there. He left to join Tinder and you know, took a little coaxing. It took a couple tries. I did a short stint at Sony PlayStation but then ultimately my friend kind of approached me and said that hey, they had an opportunity for a DevOps engineer to join.

So I joined in September 2017 and right at the point when pretty much Tinder was, I mean it had already kind of taken off but like it really was starting to take off at that point due to some new features we'd introduced in I think the summer of 2017. And so I joined at like a really interesting time where scale was a challenge. You know, they had, you know, everyone knew about Tinder, they were trying to use it and they were having issues related to resilience and scalability. So it was a great learning opportunity. And yeah.

Matt Pacheco
So would you say your experience with some of the previous platforms like the file sharing and the gaming website informed your approach to your work at Tinder?

Chris O'Brien
Oh, 100%. So, FileFront was not the only file sharing website that I worked on. I actually worked at a company called MediaFire. So MediaFire at one point was a top 30 website in the world. Just give you an idea of the scale. So that was all on prem though there was no cloud-based infrastructure. But that being said, I, you know, I think I did kind of learn some interesting, like networking, you know, Skills, troubleshooting skills that ultimately were like really helpful at Tinder for sure.

05:54 – Implementing a Reliable, Scalable Cloud Infrastructure

Matt Pacheco
So let's talk a little bit about your experience with founding Tinder's cloud infrastructure team. So, so you did that. Can you share the story about that and some of your initial challenges when you did that?

Chris O'Brien
Yeah, for sure. So I, yeah, I joined as an IC DevOps engineer for Tinder and I joined a kind of traditionally systems engineering team. I mean they were, they had come from those backgrounds and ultimately what we decided is that we wanted to move towards more of like a DevOps mindset where were going to try to move to a self-service model and empower other engineering teams kind of deploy their services and infrastructure. So once we kind of set that mission, we also had some other technologies in mind as well. So the CI CD process at that point weren't particularly robust. Right. So we made the decision to kind of build our own CICD pipeline and additionally Kubernetes was just becoming fairly well known and fairly popular at that time and we felt like that could be a really good opportunity.

Not in terms of like not only in terms of scalability and reliability but also in terms of cost because at that time were basically just deploying a single service to a single EC2 instance which was not necessarily the best use of resources. So once we kind of had that mandate like that, were going to move to the self-service CICD model. We're going to move everything to Kubernetes. We kind of use that momentum to build the team out. So we started as a team of three, quickly hired a fourth and then over the years picked up through hiring, through kind of merging with other teams. We ultimately have ended up being a much larger organization and it seems like.

Matt Pacheco
The app is constantly growing, and you have more interactions potentially because you're a global company. How do you, with your cloud infrastructure, approach reliability when dealing with these billions of potential interactions on a daily basis?

Chris O'Brien
Yeah, we're fortunate enough to be on AWS because AWS does solve a lot of the on-prem scaling challenges that I used to deal with when I wasn't working in kind of the cloud native infrastructure. So that obviously helps quite a bit in terms of automatically provisioning compute, and you know, storage for our key value stores and you know, giving us some elastic capacity for like you know, our relational databases or caching tier. But on top of that I think how we more or less approach it is taking advantage of you know, some of these other like really well-known open source technologies. I mentioned Kubernetes. Right.

We do rely quite a bit on Kubernetes to basically schedule all of the workloads at Tinder, pretty much all the workload that Tinder, so everything is containerized, allows us to kind of, you know, on demand take advantage of the elastic compute that AWS is providing us and put containers on those instances quickly. And, and additionally we have, you know, envoy to allow us to do kind of some advanced routing logic in terms of networking. Right. So we, you know, have the ability to shape traffic depending on requests that customers are making as well. Like if a particular service is having an issue or a particular container is having an issue, we can route traffic away from it. So we kind of combine, you know, those like under the hood fundamental technologies are like kind of fundamental platform with really strong monitoring and observability.

So we keep an eye on how the app is performing at the service level as well as, and teams are on call for their respective services. And then additionally like my team is responsible for taking a look at like the top line, you know, at the edge. What are we seeing in terms of error rate and latency and some of those key metrics.

Matt Pacheco
That's really interesting. You actually just mentioned the edge and I was going to ask you about the edge and what role does the edge computing play in your infrastructure strategy?

Chris O'Brien
Yeah, so we actually took kind of an interesting approach. So we don't actually do any computation at the edge. All the computation takes place in our, on the back end where we're hosted in on AWS. That being said, we took advantage of, I think a few years ago we took advantage. Well, definitely a few years ago we took advantage of putting everything, including API requests behind CDM. And because we are a global business, what we wanted to do was terminate that connection closer to the end user. Specifically tls, we wanted to like terminate HTTPs close to the user so that they wouldn't necessarily have all of that round trip latency getting all the way back to the US and then maybe going all the way back to Europe or Asia or whatever the case may be.

So we don't necessarily do any computing at the edge, but what we do instead is try to take advantage of geographic, like geographically close points of presence so that we can kind of speed up some of the initial connection to Tinder. And then we leverage, we leverage Amazon's backbone to basically handle all of the connection, the connection back to the origin.

Matt Pacheco
You have a global app, billions of interactions, you're delivering new features often. How do you balance that more rapid feature development with infrastructure kind of stability? Like how do you ensure that?

Chris O'Brien
Because yeah, under the hood everything's elastic, right? So you know, we, as I mentioned in like kind of the top of the call, we, everything is self-service model, right? So we expect backend teams if they're developing a new feature, or really any teams if they're developing a new feature that they are deploying and managing the infrastructure for their specific service. And based on those requests, we will, you know, the, the capacity will already be there or we will automatically provision that capacity. So we have gotten out of the, I think like the fairly legacy approach of like, you know, getting a ticket, having a new service that needs to be onboarded, spinning up the resources, deploying it there instead.

Like we have automated all of that to the point where if I have a new service that I want to deploy today, of course it has to go through security review and you know, product review and executive review and all these different functions, right? To make sure that this is something that we actually want to deploy. But once we've made that decision to deploy, I define my schema, I deploy that and it's automatically provisioned, right? How we mitigate that kind of fast iteration or the fast iteration combined with resiliency is we again keep a very close eye at the service level and also at the edge in terms of like what customers are seeing in terms of error rate, latency, etc.

And we also take advantage of things like, you know, canary analysis, like automated canary analysis, where the new service will get pushed in production and if there's an issue, it will get immediately rolled back. And we're in the process of kind of like, you know, even furthering that anomaly detection. So that anytime there is an immediate issue with a particular, with a new service, it will throw a bunch of alerts and create an incident for our teams.

13:47 – Storage and Protection for Data in Online Dating

Matt Pacheco
So I'd like to talk a little bit more about data and potentially talk a little bit about AI as well, since that's the hottest topic in the cloud world. I keep hearing about it and everyone's doing something a little different with it. But first I want to ask you about data. We often hear the conversation, do we do data warehouses or data lakes? There's always a comparison between the two. And we've heard it from, from others that there's the concept of lake houses. So it's really interesting. Can you talk a little bit about what Tinder how you approach data in that way?

Chris O'Brien
Yeah, for sure. So we have a data lake and we basically have different tiers of data that exist in that data lake with different SLAs. The kind of two conceptual differences, I think the biggest, probably the two largest conceptual differences between all those different types of data is that one is just raw, which is unenriched, just basically comes in from the client, goes straight to the data lake and then you have kind of a roll up tier or even before that you might have an enrichment layer where essentially like we're taking those raw events in and potentially enriching them with you know, additional data from other sources or creating a roll up where we're aggregating a bunch of different sources together.

So each of those different tables have you know, obviously bronze silver tier, gold level SLAs and monitoring and availability and we, and yeah, kind of depending on the use case will determine what, what data you'll use. Right. So you have a lot of like, you know, like internal stakeholders who need business analytics data querying the gold tier, which is kind of our roll ups or data cubes. And then you might have engineering, you know, customers that are potentially leveraging the raw layer for some of their, some of their use cases. But it's basically one big data lake. And we do, yeah, I mean we're talking many petabytes in size and ingesting terabytes, you know, tens of terabytes a day in analytics events.

Matt Pacheco
Yeah, I can only imagine. Is there an opportunity in there or have you already started using machine learning, AI as it relates to data?

Chris O'Brien
Yeah, so we have machine learning teams at Tinder. They're focused in several different areas around, you know, using AI and large language models to improve the customer experience. So they take advantage of the data that's in the data lake to kind of train their models and gear some of those predictions or text generation towards, you know, specific, you know, taking advantage of that data. So I think I answered your question. I mean tell me if I, if you want a little bit more detail there.

Matt Pacheco
A little follow up question is. All right, so you answered 1. Teams are using that data for, with machine learning. Do you use AI and large language models for other initiatives? So like for cloud infrastructure as well?

Chris O'Brien
We are in, yeah, we are in the process. So we basically. Yes. So right now most of our focus is on anomaly detection. So we want to be able to create incidents and let teams know if we have detected anomaly in service Level metrics or maybe in date data. Whether like the data be malformed or we're not getting as many events as we used to. So that's really where most of our focus is. I would say on the generative AI side that's predominantly I think geared towards the customer use cases.

Matt Pacheco
Yeah, I would, that's often something we hear is that the machine learning and AI is potentially for anomalies. We'll talk a little bit in a few about security use cases as well. And then yeah, the, a lot of companies are using it for product enhancements, chatbots and, and stuff like that. So I was curious, I figured I'd ask, I mean while we're talking about security, we can go into that area as well.

Before I ask the AI question about security, I'm curious about data loss prevention on such a massive scale. Can you talk about, because that's a lot of data you guys manage, you were saying terabytes per day, petabytes in total. That's, that's a whole lot of data for people around the whole world. There's a lot of compliance things to worry about depending on the region. So I'm curious about your approach to data loss prevention.

Chris O'Brien
Yeah, data loss prevention. So we again we're fortunate enough to have, be able to leverage a number of different AWS technologies. Right. So for all the AWS managed services, they all have backup policies that we can enable for any like really business critical data. We actually have our own process by which we're copying that data and shipping it to a different region. Right. For disaster recovery purposes. Additionally, you know, we use IAM role based management for querying this data so that provide, you know, obviously making sure that only the service can kind of read and write some of this data and it kind of limits some of the blast radius if something were to go wrong.

And then that kind of ties in with the other two strategies which is AWS as backups and then we have some of this data available in other regions.

19:39 – Having High Standards for Privacy and Compliance

Matt Pacheco
So you got into a little bit about security with IAM and a little bit of that. What's kind of your approach for security across a global cloud infrastructure? What's important to you to include in your strategy to protect all of that?

Chris O'Brien
Yeah, great question. I mean you touched on some of the regulatory compliance, right? So we work with different regulatory stakeholders to make sure that we're compliant depending on the country. So I think that's a big part of our focus and is A huge priority for us. Beyond that, we do have the advantage of being hosted in a single region. Right. So we can, regardless of the global footprint, we can basically apply easily, you know, relatively easily apply the same kind of security process and enforcement to that specific region. And we mentioned IAM the security team is great about doing pen testing as well as they have red teams that are ultimately trying to find bugs or you know, find issues before they surface to people who are more interested in doing damage rather than helping us out.

And we also kind of do regular, we have all sorts of monitoring and alerting in terms of like if something is accessing data that it shouldn't or you know, or we see requests that seem malicious or malformed, you know, we get alerts on those as well.

Matt Pacheco
That's interesting. You talked about data privacy and compliance requirements a little I'm curious for a platform like Tinder, what are some of the more unique security challenges? You talked a little bit about some of the threats you face, but are there unique security challenges for the kind of platform or is there any interesting things to consider?

Chris O'Brien
I mean, I think like when you're a globally, like a well-known brand, like a global brand, I think that poses, you know, become, then you become a target. So I think that poses its own kind of unique challenge. But I think like probably the unique challenge that we have is what I mentioned earlier is that we basically have a self-service model where teams kind of can deploy and manage their own services and infrastructure. That of course then becomes very dependent on the teams being aware of some of these security concerns on a team level and being able to make sure that they're following the various different guidelines or best practices that we expect teams.

So that how we mitigate that is of course through training, making sure really well aligned with the teams on what's expected of them as well as just trying to provide some guard rails related to like, you know, if you're going to deploy for example, like a new route that's publicly accessible, not just internal endpoint, but actually an external endpoint, then that has kind of a number of automated steps by which it needs to go through before it's deployed.

Matt Pacheco
Any rising security trends that concern you?

Chris O'Brien
I mean, I think like, you know, something that definitely concerns me is you know, how generative AI could be used to exploit different things. Right. So obviously there's now opportunities for, you know, entire code bases to potentially be ingested into a model and being analyzed for some sort of vulnerability or some sort of weakness. You also have, you know, the opportunity to use AI to write a bunch of different requests or maybe write code quickly to allow you to exploit a bunch of different endpoints. Yeah, I mean, I think like, I mean, it's just like the same efficiencies that you and I are getting from AI, from using ChatGPT or Llama or whatever sort of model that we're using Claude, as part of some of these code editors, like you're now, you know, you're getting those efficiencies.

Then I think these attackers are getting these efficiencies as well. On the trust and safety side, you know, it's also, now they can, now that actors can generate, you know, like use generative AI to write conversations. Right. They can look a lot more real to our customers. Right. So we, it's not as easy, it could potentially not be as easy for us to detect, you know, bad actors or bots on the platform in the past. You know, something you can, you know, I would say maybe not in every case, but you can read through these chats and understand, okay, that's probably a bot. You know, if you're maybe talking with a well-trained model, you may not necessarily get that impression. Right.

And our models that are supposed to detect whether or not it's a bot may not be easily to be able to pick up on that signal.

Matt Pacheco
Yeah, it seems like a constant arms race of who can have the better AI bots. Very interesting. And from my perspective, from my putting on my consumer hat, I've seen a lot of the phishing attempts becoming much better with the use of ChatGPT. No more spelling errors that you get years ago and it's becoming harder to do it. So I can imagine from a platform perspective, people using the platform and talking these AIs, it's probably very easy to get fooled. So, so next I would like to talk about a rising issue in the entire industry, all of it. And that's teams, that, that's people, that's talent. As we know, there's a shortage in security and cloud, a shortage of people who have the skills, who can come in with the skills that are needed. Infrastructure, and a lot of businesses are struggling with that.

25:28 – Experts in Matchmaking: How Tinder Attracts Top Talent

Matt Pacheco
The talent gets quickly scouted and picked up. How do you, how do you attract and retain good cloud talent? Like, how do you overcome that challenge?

Chris O'Brien
Yeah, that's a great question. So we have, we've been struggling with this for several years now. And I think what we have had to do is kind of twofold. One is expect different skills. So I think the people like myself who come from compiling Linux kernel or doing a bunch of systems administrations or systems engineering are becoming fewer and far between because most companies are moving cloud native, and you don't have to necessarily deal with those issues or things. Right. There's a layer of abstraction that you're working with that I didn't necessarily have the opportunity to do that early on in my career. So we, and a lot of those people exist, right, but they're working in existing companies or you know, they're well established at bigger companies or startups or whatever the case may be.

So we've had to basically shy away from recruiting for those specific skill sets and instead try to identify who would be a fit from some of the intangible skills. And I'll talk about that in a second. And also maybe just like pure development skills, right. So people who are, you know, proficient software engineers or maybe come from a DevOps background with maybe not as much experience in the system space and we prioritize hiring them and maybe even at lower levels. Right. Like we're talking junior mid-level engineers. We've tried to index on people who are like really hungry, eager to learn, curious. And what we've done is kind of created the system within the team where we have kind of a tier of like senior experienced people that are mentoring the less senior, lesser experienced people and then that kind of filters on down.

So we kind of have this new generation of senior engineers that are constantly coming throughout the team that will mentor people that we're hiring today.

Matt Pacheco
That's interesting. And then you manage a really large team as well. How do you structure and organize such a large team to run efficiently?

Chris O'Brien
Yeah, we've taken the approach of having sub teams. So we're talking about three to five people who are experts in their particular space, could be CI CD for example, and asking them to be autonomous. Right. So they, of course we give them kind of a high-level objective that rolls up into the wider organization vision, but it's on them to execute. Right. So they make the technology decisions, they select the technology, they do the implementation and it ends up kind of being a bunch of autonomous sub teams where it's kind of like a we push that responsibility as far down as we can, which lets us, I think, scale kind of the technical leadership. And by making sure that we have some of the senior people I just mentioned in those different groups.

28:48 – Planning a Future with the Cloud

Matt Pacheco
Great. So I'm going to ask you some questions about the future. Where do you see cloud infrastructure in the world of cloud evolving for social platforms like yours within the next few years?

Chris O'Brien
Makes sense. I mean I think like the big push is going to be related to GPU based infrastructure, you know, I mean pretty much you're going to see more and more social platforms trying to take advantage, even Tinder, trying to take advantage of AI and being able to do inferencing very fast. Right. So right now we have a lot of CPU like traditional compute CPU based infrastructure and I think that over the next several years you're going to start to see more capacity and infrastructure management challenges arise from basically managing these huge GPU pools. And I'd imagine that companies like AWS and Microsoft are going to try to create some platforms that kind of bundle that all together. Right. So you don't have to deal with the underlying infrastructure AWS already has.

And I'd imagine there's going to be more investment and probably teams moving more toward those managed services for GPUs, GPU infrastructure.

Matt Pacheco
Very interesting. Hearing a lot about GPU infrastructure in all these conversations is not the first time someone's bought it up. So it's really interesting to see that trend start to take hold and the big players in the cloud space adjust to that. In a similar, in a similar vein, what emerging technologies are you personally most excited about in your world?

Chris O'Brien
Oh yeah, great question. I mean I, I mean again this is going to be all AI driven so I mean I super interested in large language models so a lot of my spare time, if I'm, you know, when I have spare time is focused on training neural networks and kind of running different experiments. I have my own set up here so I don't have to worry too much about the GPU infrastructure that we talked about. And I think the other thing that's really cool is like for example these IDEs that are well integrated with LLMs. Right. So we're talking about Cursor and GitHub Copilot.

Like that is a super interesting space because I mean I think you have, you know, industry leaders like even Mark Zuckerberg is talking about how, you know, maybe mid-level engineers will be replaced over time or at least maybe in the short term by LLMs writing a lot of this code for them. And I guess I could see, I could definitely see that being a possibility. I mean these models still require oversight and inspection to make sure they're doing what you want them to do, but they can, given the right prompts and maybe like a little bit of tuning, write some pretty good code. So I've been experimenting quite a bit with that.

Matt Pacheco
Yeah, I've heard of some of those tools. Someone on the podcast has talked about Devin that tool in the past two. So it's really interesting space. And I'm very curious your thoughts about the potential regulation coming to the AI space as well. What do you think about that? I know Europe's probably far ahead of the United States, but. Curious.

Chris O'Brien
Yeah, I don't necessarily know that I'm all that well informed. I have heard about some of the California regulation that's being introduced or maybe has already been voted on. I've heard about some of the European stuff, but I'm not super up to speed on that. I will say that it's like, yeah, I mean, there's. I'm caught between this mix of like, you know, someone who, you know, myself, who's an innovator who just wants to be able to innovate freely. I'm sure a number of other companies want to as well. But also the risks that are potentially posed by AI like, probably do need some regulatory oversight. You know, my wife is a writer. She's a creative writer. She works in television, film, comic books. Right.

So her industry is extremely interested in maybe preventing the use of AI to replace them, essentially, or even using their work to train models, whether it be scripts or for movies or television shows or whatever. So I think it's probably needed. But there's a part of me that more or less wants to be able to innovate freely. And I'm sure, like I said, I'm sure other people feel the same way.

Matt Pacheco
Yeah, I think just users have to be maybe educated on using them. You don't want to stick your whole financial report or something in there before you've actually announced it, because who owns that data? Who knows what's in there? And it's just an interesting conundrum sometimes, and I'm very interested in seeing how it's regulated in the future and then how companies are going to be impacted by those regulations and what steps need to be taken. But I guess it's definitely a wait and see.

Chris O'Brien
Yeah, I mean, I think some of the regulation is, is necessary. I mean, you know, the stuff we're doing related to data lifecycle policy I think is needed. I mean, I think customers have the right to own their data. If they delete their account, we shouldn't keep their data or it should be anonymized. Like, I think these are really good regulations. But I will say, speaking from experience, someone who works in the data space, this creates a lot of engineering work for our teams and also a number of other stakeholder teams. Legal privacy products.

Matt Pacheco
Absolutely. So I think I have, like, one more question for you. We talked about a lot. It was great. So we have a lot of listeners varying from IT leaders to people in the field with their hands on. What is some advice you can give? We can start with the leaders to leadership on thinking about their cloud infrastructure and possibly a global operation. What are some things they should think about or advice you can give to those leaders?

Chris O'Brien
That's a great question. So I think that my advice, generally speaking to any leader, is that, you know, we're here to empower our teams. Right. And help them kind of develop and grow. So I, I think in that same vein, creating learning opportunities for yourself, whether it be in technology or in leadership practice and kind of instilling that within your teams, I think is essential. I think it's been essential to the success we've had at Tinder. And I really would encourage all leaders to constantly be encouraging themselves as well as their teams to kind of learn and grow and to kind of challenge some of the status quo in terms of how they're doing things, whether it be how they're managing a team or how they're approaching leadership within their group or even just like what technologies they're using. I think there's never.

Things are constantly evolving and we should always be trying to kind of learn and grow.

Matt Pacheco
Excellent. And then for the other audience who's potentially getting into the space as well, what are some things or advice you can give to them as far as, like, learning or skills? They should have to break into the cloud industry.

Chris O'Brien
Yeah, for sure. I mean, I think one bit of advice is that you are better at what you do than you think. I think that a number of people struggle with insecurity, even myself. And I think the reality is that we're very tough critics of ourselves. We're very hard on ourselves. Right. So I think it's important to keep in mind that you're. You're further along than you think, whether it be in your career or your learning or studies. Beyond that, I think it's really important to be curious. Right. Be curious in your space. Right. So if you are trying to, if you're in school or you're entry level in particular, team or technology space, learn everything you possibly can about that space. Ask, be really annoying. Ask lots of questions.

My mentor told me and it has resonated with me to this day that smart people ask questions, they say. And they also say that they don't know. You don't get any points by knowing everything. Right. And having that humility. Humility and curiosity when it comes to what you're doing as well as what other people are doing, I think will go a long way.

Matt Pacheco
That's actually amazing advice. That's awesome. And that can apply to everyone. So I appreciate the advice and the knowledge you've shared with us today at a great time. I didn't realize how fast this went. It's really enjoyable and it goes really fast when you're enjoying the conversation. So we really appreciate you coming on here and telling us a little about yourself and a little about Tinder and some of the trends you're seeing in the space.

Chris O'Brien
Thank you. I really appreciate the opportunity. Yeah, it went by fast for sure. And yeah, maybe, maybe you have me back on. We can talk about something else, you know, related to Tinder or whatever the case may be.

Matt Pacheco
Yeah, there's definitely emerging trends constantly popping up in the cloud world. So we'd love to have you back. Awesome. Well, thank you and thank you for our listeners. We appreciate you watching and listening in to the TierPoint’s Cloud Currents podcast. Stay tuned for the next episode. You can find us on all major platforms, including YouTube, and we will see you next time.