Demetrios Brinkmann has built his community in a really interesting way.
For starters, he operates in a really niche niche: machine learning operations, or MLOps. Machine learning is the branch of artificial intelligence that's involved in creating systems that can learn from data, identify patterns, and make decisions with little human intervention.
You know how Netflix selects the next show for you to binge-watch? Yeah, that's machine learning at work.
MLOps is all about how machine learning is put into production so people and organizations can realize business value from it.
And Demetrios has created a niche community for MLOps professionals in a clever and ultimately successful way.
He's managed to foster a sustainable space where people are excited about connecting and sharing value, where sponsors have a place (but spam is kept to the sidelines!), and—oh yeah—it's free to join.
In today's episode, you'll learn how Demetrios balances multiple tools to pull his community together without creating overwhelm or confusion. You'll get a taste of the nifty onboarding sequence Demetrios created that helps keep spammers in check. And you'll discover how he landed on just the right monetization scheme to keep the community humming.
Demetrios Brinkmann is one of the main organizers of the MLOps Community and currently resides in a small town outside Frankfurt, Germany. He is an avid traveler who taught English as a second language to see the world and learn about new cultures. Brinkmann fell into the Machine Learning Operations world, and since, has interviewed the leading names around MLOps, Data Science, and Machine Learning. Since diving into the nitty-gritty of ML Operations he felt a strong calling to explore the ethical issues surrounding AI/ML. When he is not conducting interviews you can find him making stone stackings with his daughter in the woods or playing the ukulele by the campfire.
In This Episode
- Demetrios's struggle to explain what he does to his neighbors and friends
- The collision of machine learning operations and community
- What the chatter is all about in MLOps Community
- How Demetrios pulled MLOps veterans into his community
- Organizing multiple platforms for a community
- How Demetrios approaches the challenge of monetizing his community
- Handling community spam
- Garnering sponsorships for an online community (and reinvesting them into the community)
- How to Talk So Little Kids Will Listen by Joanna Faber and Julie King [Amazon affiliate link]
- Principles for Dealing with the Changing World Order by Ray Dalio [Amazon affiliate link]
The CX 028: Do Things That Don’t Scale with Demetrios Brinkmann of MLOps Community
Demetrios Brinkmann: A lot of the sponsors that have come are already in the community and they've realized a ton of value from the community because they're hearing all these conversations that they wouldn't necessarily hear.
Tony Bacigalupo: Hey, Hey, Hey, what's up, everybody. It's the Community Experience podcast episode 28. I'm Tony Bacigalupo.
Jillian Benbow: And I am Jillian Benbow.
Tony: And today we are going to be talking with Demetrios Brinkmann, who runs a community called MLOps.community. And if you don't know what that means, you're going to find out. But it doesn't matter as much as how he built his amazing online community for professionals who are all in a very specific niche industry, how he did it in a way that was very clever, getting people excited, getting people talking to each other, sharing value online, how he made it free, but still kept it high quality, kept out the spam, and how he was able to secure some great sponsorship and really find a sustainable model. That's a whole bunch of stuff.
Tony: Jill, what are you excited about? What are you looking forward to in this conversation?
Jillian: All of it, but especially how to incorporate authentic sponsorship opportunities in your community.
Tony: Let's do this. Let's get into the conversation with Demetrios Brinkmann of MLOps.
Jillian: Let's jump in.
Tony: Demetrios, welcome to the program. So great to have you here. Really excited to chat with you.
Demetrios: Hey, everybody. What's going on? I'm so stoked to be here.
Jillian: We are happy to have you, and you're joining us from Germany. You're in Frankfurt, correct?
Demetrios: Yeah. I'm actually in a little village outside of Frankfurt, which is different for me because there's about 100 people that live in this village.
Jillian: Wow. I can only imagine when you tell them what you do for a living between the language barrier and then just like the small town vibe, do they just smile and nod and they're like, "That's nice. He's a drug dealer."
Demetrios: Yeah, basically.
Jillian: It's a front.
Demetrios: Yeah. People ask me, like they still have trouble figuring it out, for sure. It's one of those things where I have to repeat it various times and then they just get confused and they say, "All right, let's talk about something else."
Jillian: I still experience that state-side with community, so I can only imagine. Anyways.
Tony: Sometimes it's like, I do computer stuff. I do stuff with people like computers.
Jillian: Internet stuff.
Demetrios: Yeah. Yeah. So it's even like one step further because I try and explain that I do stuff with community and then they say, "So what's your community about?" And that's when I really get the glazed eyes, if they weren't already glazed, because then I say, "Well, do you know what machine learning operations is?" They generally like 90% of the time say, "No." And then I'm like, "Do you know what machine learning is?" "No." "Do you know what AI is?" And they're like, "Oh my God, it's going to take over the world." So then I'm like, "Yeah, all right, something around that." And they're like, "Okay."
Jillian: Well, that's the perfect segue to tell us what you do do in machine learning and community, how do those two collide. What do you do, Demetrios, in your little village in Germany?
Tony: If you can, Demetrios, maybe for the sake of our human learning, can you give us your best distillation of your explanation of what machine learning is?
Demetrios: There's machine learning and then what I have a community around is called MLOps. So it's the operational side of machine learning. Machine learning is basically like when you go to Netflix or when you go to Amazon and you buy something or you watch a video, then you get recommended a bunch of other videos that you should watch or a bunch of other stuff you should watch. That is one form of machine learning. And it's not hard coded into the Amazon system because it takes all this data from all these different user profiles that are probably a little similar to you, and it says, well, these people liked these other shows for the Netflix example, and then it will feed you some of these shows and say, you might want to watch this, and it recommends you these shows. That's one form of it.
You also have other forms where it's being used in fraud detection, it's being used in robotics. All of that stuff that you see with like Boston Dynamics, all of those viral videos, they're using machine learning in that. There's all kinds of different machine learning applications and use cases that are happening right now. You've also got computer vision. I mean, anyone who has not been living under a rock has probably heard about Tesla and the self-driving cars, and that is using machine learning. And that's a very different kind of use case of machine learning. But machine learning itself is when you have those algorithms and you're trying to figure out how to make those algorithms make decisions for you instead of having a human make these decisions. And then the operations part is all right, let's imagine we've got this algorithm over here, but now we need to put it into production and start realizing business value out of it.
So instead of it just being on a data scientist laptop, where they created the algorithm and it worked fine, now we need to go serve it up to millions of people and really scale it to whatever Netflix size is or Uber or all of these things that use machine learning. Ubers case for example is how do they know how long it's going to take the driver to be at where you are when you order a car, right? Or how do they recommend different food that they want you to eat? So the operations part is really difficult because you've got the machine learning algorithm that is doing its thing, but then it needs to be making these predictions and it needs to be served up to all of the people that are users of the platforms, whether that's Netflix or Uber or whatever may have you. Or if it's in the case of the Tesla car, it needs to make sure that it breaks at the right time or it sees the human that is crossing the road.
So when you want to put it into production and start realizing business value out of machine learning, that's when you come into the space of machine learning operations. As for that other question of what I do, I have a community that I just kind of fell into being the organizer of, I guess, is what I would call myself, and it is all around machine learning operations. So it's a super niche field of the machine learning space. It's not exactly like looking at the algorithms, it's more looking at how to put those algorithms into play.
Jillian: I mean, I'm all for some of these things taking over, like yes, Netflix, I would like you to know what I want to watch next. Yes, please. I'm here for that.
Demetrios: Just autoplay it? Yeah, please.
Tony: So these tools are very technologically powerful and these folks who work in Ops kind of sit at the middle space between that technology, which can go a lot of different directions, can be you misused or just misunderstood, and then the people who are actually going to be using it or dealing with how it works. So Tesla being a very kind of like mainstream example, how that car behaves when it's driving itself and when things maybe go wrong and there's all these things. So these are people who can connect over the common ground they have sitting at that intersection and talk about maybe how to communicate better about it, and maybe how to better use these tools and talk to people about them. Is that the kind of stuff that people discuss?
Demetrios: Yeah, exactly. So right now, there's a lot of, it's like shrouded in mystery because machine learning is such a new field and machine learning operations is still even newer than just machine learning. And so realizing business value from machine learning is something that you see lots of the big companies doing, like the Ubers and Netflix, because they have the headcount, they have the resources to throw out these problems. But now in the last like two years, I would say even just like in the last year, you've seen the smaller companies start to realize value also. Because of the maturity of the space, more and more companies are able to get into this space, and they're able to bridge that gap to where it's not only the Netflix and Amazons that are realizing value from machine learning. There's new startups that are able to really find their niche and find how they can take advantage of machine learning.
And so that's where we sit at. We're sitting at that cutting edge area where it's like, people want to come, and since it is new, there's not a lot of information out there about it. And we've got this community where you have luckily, and I don't really know how this happened, I can give you a few tips, but we've got some battle harden veterans in the community that talk about their war stories openly. And so that is very useful for the others who are just starting off on the path, and they've been tasked by their boss to put this machine learning model into production.
Tony: How did you get those battle tested folks in the room?
Demetrios: That is something I still I'm not fully sure of. One thing that I think I really tried to do at the beginning of the community, like formation, was ask people that I knew were really smart to not only join the community, they were already in the community maybe, and they showed in a few answers that, wow, they know what they're talking about. So it was doing things that don't scale, reaching out to those people, having one-on-one conversations with them, and then saying, "You know what would be super useful is if you can ask more questions or you can share how you are looking at things in Slack." Because we have a few, and we could probably get into this, but we have a few different areas where we congregate, but one of them is our Slack workspace.
And so I just asked these people that I felt like were some of these veterans. I said, "If you see a question that doesn't have an answer, do you think you can chime in and give your two cents? If you are looking at something or you find a really cool article, can you share it in here?" Just trying to get more movement, so it wasn't only me sharing things in the community. And it felt like there was more people.
Tony: It's so funny because you said you didn't know, and then you gave the perfect textbook answer of exactly how you get those kinds of people in the community.
Jillian: It sort of work through it. It's fine.
Tony: And that does roll into another question I had, which is how you do congregate, what are the different forms of engagement that you have in your community?
Demetrios: Yeah. Mainly I think where we probably have the most numbers is in Slack. Right now we just hit around 8,000 people. In the beginning, I can talk about the whole story of it, but the Slack is where we have the most people. We've also got all kinds of different stuff that we're doing, right? Like YouTube meetups or just meetups in general, virtual meetups, and we recently started doing live, in-person meetups, then we've got podcasts around this, we've got blogs. But those I consider more, and you all probably know this better than I do, it's like those are audience building things. Those aren't necessarily community things, unless you really try to source these different guests or these blog posts from the community. And that's what we've been trying to do.
So to make it not just like me pumping out content to the rest of the community, it's like the community is sharing content with each other and-
Jillian: The magic sauce right there.
Demetrios: That's it. That's what you want to hopefully get to. Yeah. In the beginning, we had Discourse like the forms and that didn't work out just because I wasn't big on forms and I don't think I gave it the love and attention it needed. I stuck to Slack and then that got more traction and I said, "Well, let's get rid of Discourse." And we didn't go with Discord just because it's so noisy and it's so crazy chaotic. I have a hard time being in there. I am part of a few other Discord. Especially in the blockchain in crypto world, most people use Discord, but I tend to find it like, oh, I get a little dizzy after being in there for too long. And so Slack felt like a good fit. We went with Slack, and that's where we generally are.
What we just started implementing actually which is super cool, and I love this tool, like huge shout out to the people at Gradual, it's an events platform, but it's way cheaper than Hopin or whatever other events platforms that you've seen. I don't know what other one, I can't remember. You all probably know a lot of great platforms, but Gradual, it's just got this slick interface, like the videos, the videos follow you around. And then they also have a lot of these cool features that most events platforms have, but they're taking more of a community, like we expect people to hang out on here more than just come for one event. So that's one other area that we're just starting to move towards. It's not going to replace Slack, but it is going to be something that we can have in parallel.
Jillian: So you use a lot of platforms for this community, and it sounds like Slack is the central hub, but then there's offshoots depending. That seems tricky.
Demetrios: Actually, just as another thing, now that you mentioned that, I just forgot yesterday, we're going to start using this other tool that I should probably give a shout out to. It's these guys called Pallet. Have you heard of them? And they actually, they are for job boards. It's like curated job boards. And that's another incredible tool that I'm really excited to use because we've been having a jobs channel in Alack and it's really busy. And the straw that broke the camels back, like I get reached out to quite a bit from recruiters or people asking, "Do you know somebody who would be perfect for this job?" Blah, blah, blah. And it's great, but I remember I can tell you this story. It might make me sound like a bad person, but this guy reached out to me and he said, "Hey, we're well funded startup, and you're in the MLOps space, and we're looking for an MLOps engineer. Can you refer anybody to us?" And this was, whatever, the recruitment lead for a company. And I said, "Okay, is there a referral fee?"
And he said, "No, we're a startup. We can't be giving any money to... Sadly, right now we don't have that kind of money." And then I read the job description, and the first line, literally the first in the job description said, "We are a startup with over $175 million in funding by Andreessen Horowitz and Sequoia and all these big names." And so I just took a screenshot of that and then sent it back to them and said, "All right, when you want to talk." But anyway, Pallet kind of solves that problem because it is a way for the community managers or the community organizers to get a piece of the action if you do make that connection.
So back to that other question that you were saying is like, how do we keep it all together because there is a lot that's going on. What we started implementing, which is really cool, and this is just a simple notion doc. We created like a community orientation guide and it's a notion page and it's like the community docs. And so you have everything that we're doing and every tool that we're using and everything that is related to something, and it just gives someone really a bird's eye view of all of the different events we have going on, if it's like the reading club or if we have round table discussions or the podcast or whatever. And it gives you links to them, it tells you what they are and where you can find out more information or who is leading the events, if you want to reach out and you want to contribute, all that kind of stuff. So it helps people be more self-organized in that way.
Jillian: I love that. Notion is the best.
Tony: Is that publicly visible or is it members only?
Demetrios: It's public everything, except for one of the pages that has potentially sensitive information that I didn't want vendors to be seeing because it's like people in the community talking about what their technology is that they're using. And I was just afraid like, oh, vendors are going to see this, because there's a constant battle between the community. Since we are a non-biased or vendor neutral community, there's always that. And I can get into all that fun stuff later, like how to reconcile sponsorship and vendors, and the moral fabric of the community and the integrity. But one of those pages has information about stuff that I considered sensitive, so you have to request access.
Jillian: I think that's smart. You got to protect your community members above all, right? So the creation of this notion database, I'm sure you built it over time just as things came up, but was the community a part of creating what would be in it? Or did you just kind of pay attention to the common questions and whatnot and realize like, oh, we should have all our events here, oh, we should have info about X, Y, Z?
Demetrios: Yeah. We had certain iterations of this and I think that's kind of the way things work with community, for sure, like you iterate so much on things because it is like a group that is making something happen as opposed to one person or like a top down style of doing it. And so whenever there's a suggestion about we should do this or we should do that, I usually start with creating a working groups channel in Slack. And whoever was part of the people that suggested what we should have or at has shown interest, I'll throw them in that channel, and then we start brainstorming about it, and we start iterating on it. And so originally, we thought, oh, well maybe we could take all of the good threads that we have in Slack and condense them down, and it'll be like a knowledge hub where it's like the best information about X, Y, Z, and we create a from page out of that.
It turns out that is a really complicated task to do because in this ever-changing environment of machine learning, this week, it's one thing, next week, it's a totally different thing, and you are just creating a lot of work for yourself if you want to try and be up to date on the newest things and keep like that page up to date. And so we just decided against it. But what we did decide and what we did see was that people were like, "Whoa, there's a meetup? I didn't even realize that was happening right now." And so after enough of that, I was like, "Wow, we need to create a place that everyone can see." There's a new newsletter. There's a meetup. There's like this reading group. There's the round tables. We've got a whole tooling comparison page. All of the different resources that we have and everything that we're doing, I want someone to come into the community and easily see things and then be able to click into what they like.
Jillian: My head's just spinning because we have a pain point in SPI Pro, the paid membership group that we run where a very common, a very common reaction when someone joins is it's like drinking from the fire hose because there's just so much going on, there's so many conversations, there's events, there's masterminds, there's book club. And it's amazing someone will be like, "There's a book club?" And I'm like, "It's been here the whole time. You've been here a year. How did you not know?" But that's because I'm in there all day every day. So I'm just thinking like, huh, we should do this because we have the information, but it's all over the place depending, right?
Demetrios: Yeah, centralizing it was huge for us. And it's exactly that, that was the pain that I felt, and that was why I realized. I took a weekend and I just grabbed everything that I could, everything that I could think of that we did, and then threw it in one place, and yeah, the results have been incredible so far.
Jillian: That's great. Yeah. I think, well, and in our situation, I could definitely throw out a skeleton version and then ask people, "What's missing?" Because I'm sure my community knows better than I do because I can't see... What is it? I can't see the forest from the trees. Like I'm too deep. I need someone else to come look and say, "What about this?" So you kind of organized this community and it's evolved and grown over time. You're in all these places, but you have this great central notion of knowledge hub so people stay connected. Is your community monetized in any way? Is it a free community for members? And then tell us about sponsorships, Demetrios, because I know that's a thing.
Demetrios: Yeah, yeah, for sure.
Jillian: Let me just say.
Demetrios: Come right out with it. Definitely.
Tony: You talk about money.
Jillian: Tell us everything.
Demetrios: Yeah. And I'm really cool with being as transparent as you want to be or as you would like me to get, because I think it is a really difficult problem that a lot of community managers or community founders face, and that's how can you monetize without destroying the community? So especially in the beginning, I thought long and hard about this and I was trying to figure out what we could do and different monetization schemes, what would be the best way to go about it, and ultimately I landed on a... So the other thing is that we've been growing like crazy and it's free for anyone to come in, which it's kind of hard because we get a lot of vendors that come in and then spam everything. And so that it was, or is... Actually, oh my God, I'm telling you about tools all the time, but we use another tool in Slack and they also has this with Discord and I saw the vendor spam, the spender van go down like tenfold when we started using this tool, it's called Campfire, right?
And it's just a onboarding guide, kind of like choose your own adventure that you can click on as you first come into the community and it'll say, what do you consider yourself? And I created the onboarding guide. And so it's like, what do you create? What do you consider yourself? A practitioner, just starting, or working at a tooling company? And depending on what they click on, then they get a tailored message and more questions, and then they can click on one and then get more questions. And so if it says, if they come in and they say, "Well, I'm a vendor," then instantly, the next message is, "You better read our docs and do not spam anything." In a lot nicer words than that, right? But being able to do that made my work go from every day deleting some kind of spam from the community to now maybe it's once a week-
Jillian: That's amazing.
Demetrios: ... which is really good. I know it's still is a lot, probably, for a lot of people.
Jillian: For a free open community, that's amazing. Onboarding is so important.
Tony: I find it interesting too. So I actually, while we're talking, I joined your Slack group and I got your-
Jillian: Oh nice.
Tony: ... automated DM from you with this exact flow. And it's interesting because it's optional. It's not like a gated thing where if I don't answer the questions, I can't post. So somebody who wants to spam still could do it, but you're saying that there's just this added layer of a nudge that gets people to self-identify maybe just to understand the rules of the road a little bit. And I think that's really, really clever.
Demetrios: The other thing that we did, and you can probably see that as you're in the community, is we have a channel that is specifically for self-promotion. We call it be shameless. So we have a be shameless channel and that is so that you can go and whatever you want to promote, you can promote it there. And so that is a lot more than most communities do. And if I do find spam, it's like, come on, I'm giving you a whole channel, just go do it in there. And that's, by far, the most active channel in the community, because everybody's got something that they want to promote.
Jillian: Nice. It's smart to have a place for it.
Tony: There is something nice about just creating the clear expectation for everybody involved that this is what's going to happen in this space. So if I, just as a regular user, see that there's a lot of activity in that space, I'm going in knowing exactly what I'm getting and maybe I'm going in there because I want to find somebody that I want to buy something from, or I want to learn about some new products, but I'm opting into that by clicking on that channel, or I can mute it, or just leave that channel if I really have no interest in it.
Demetrios: That's it. That is totally the idea there. And sorry, I got a little sidetracked when we were talking about sponsorship and that money aspect.
Tony: It's okay. We weren't going to forget.
Demetrios: So basically, what I ended up finding was... And what was hard for me in the beginning was how can I monetize something to the level that I want to monetize it at, if we've only got like a 1,000 people in Slack, or we've only got like a few 100 listens on our podcasts and all of that. And it's very small scale. Especially at the beginning of last year, basically a year ago, we were at like 1,500, 2,000 people in Slack, and maybe getting an average of 200 views on YouTube and 200 listens in podcast land. And so what happened though, is that because this niche is so new and there are hardly any... So there's no other communities in this space, but there's hardly anywhere for all of these companies that just got a ton of VC money to dump all of that cash. I was able to start asking for sponsorship and people were into it.
Again, it was a lot of iterating and luckily enough, I have a day job and the community is not how I make money, and I enjoy my day job. And so all of the money that we make from sponsorships, which at this point is quite a bit, we just reinvest into the community and we do cool stuff with it. And I can talk about some of the cool stuff we do, but the idea there, what we landed on in the beginning of last year, I was given away so much in those first iterations, some of these companies that said, "No," now they're coming back and they get like a fifth of what they used to be getting. And now that's how it works because I iterate on it so much, and I realized, wow, I don't want to give away five podcast appearances and three joint blog posts and this, but that's what I was doing in the beginning, and I realized that's like too much of a commitment.
And what I ended up landing on was a yearly package, just so that a, I didn't have to always go out and find more sponsors for the next what month and try and bring in money. It's like, let's do a nice, big sponsorship package and it's going to be for a year, and these are the deliverables that I'm going to give you, and then we go from there. And so we started doing that and we got four sponsors in the beginning. I don't know if you want me to talk about numbers or anything.
Jillian: I would be interested though, to just hear like, and I think you're getting to this, but what do they get and what kind of community access, and that kind of stuff.
Demetrios: So there was a long time that we were thinking about like, oh, should we offer some kind of opt-in so that when you up for the newsletter, I'm okay with receiving sponsored emails or something, or sponsored posts or sharing my data with vendors or sponsors. Eventually that just felt sleazy in a way. It just felt like it wasn't aligned. And again, it goes back to protecting the integrity of the community. And so we didn't go for that. What they do get though, what the sponsors get, I should, say is, and now I can try and pull up the package, but it's something along the lines of you get a podcast appearance, and now our numbers have grown too, but even back a year ago, it was like, you get a podcast appearance, you get, and we have two different packages, like a silver and a gold kind of thing.
So there's a few blog posts, a few sponsor blurbs in the newsletter, and then your logo on the website and you can clip through. And then occasional, if sponsorship comes up in podcasts randomly, we'll give a shout out to the sponsors and make sure that we say things like thank you. And I try to just over deliver. So I wrote a blog post at the end of last year saying, "Here's what happened in the community over the last year." And I made sure to be like, "We've had these sponsors, we went from zero sponsors to these sponsors and we can't thank them enough because they are helping this community grow, and they're helping support this community. They took a chance on us and we are so thankful and so grateful for them to do that." Because in a way, it was a little bit of a risk to sponsor a community and it's not like we're doing a webinar for them and then they get all the leads.
And so that is another way that you can go. We just chose not to go that way. And we really don't share any data with the companies that sponsor. So all of the data is kept on like in a locked vault and the companies just get the brand awareness. It's like, we went for brand awareness as opposed to lead generation, if that makes sense, because I think what it ultimately boiled down to is like, these are the two things that we can offer sponsors, is like brand awareness or lead generation. And if we're going to be selling leads, then a, you got to like put a price on leads, and it's also, again, I'm not so excited about that. And just as a little bit of a side note, I came from sales. So I was the one, like I was tasked with going out there and getting leads. I used to be the one that was spamming communities. And so that's why I know all too well how that works.
Jillian: Do do the sponsors, are they in your community as members or are they posting?
Demetrios: Yeah. Yeah. And a lot of them are awesome. A lot of them are some of the best community members that we have. And some of them don't even really affiliate with their company. They're just there because most people who have created a MLOps company are the ones that were ahead of the curve and they've done this for a long time and then they saw what was needed. And so they created a company around that. And so they're able to answer some of these really hard questions that you're getting from the people who are going through them and they've already gone through them so they can answer them. What we really tried to instill this is a constant battle is just like, where's the line where it's answering a question and it's giving the answer or is it just spamming in a thread where you're like, "Oh, well, if you want to do that, you should just use our tool."
And so we had to come to this place where we said, "You know what, look, you can't talk about your tools." And this is something that is a really hard question in my mind because the field is so new and there's so many tools out there and maybe there is a tool that will really be perfect for this use case and this need and this problem, but then as soon as one vendor says, "Oh, our tool does that. It solves that exact problem. You should talk to us," then the other 50 or 100 vendors, they're all going to come on that thread and say the same thing. Even if it isn't perfect for it, it's just like a free for all. So we saw that happening a lot and I had to just cut it off and nip it at the bud and say, "Look, you can't talk about your tools. Sorry. If you want to answer the question, that's great. Just try and leave your product out of it."
Jillian: Yeah, it's always hard in those relationships, right? Because you have to, it's like, yes, I would like your money in exchange for X, Y, Z that you get as a sponsor and yes, you can join the community. However, you have to play by these rules. And some companies may not align with that. Although it sounds like fortunately in this very niche community, the people who are using those accounts that are technically like a sponsor account, like they are also just there to interact because it is that unique group where people can talk shop and really get into it and have cool conversations.
Demetrios: A lot of the sponsors that have come are already in the community and they've realized a ton of value from the community because they're hearing all these conversations that they wouldn't necessarily hear. And it's like doing market research, it's like doing product research in a way, seeing what people are having problems with, with other tools or seeing where people are getting stuck, and maybe they can add a feature to their tool that will help in that regard. And so a lot of these sponsors say, "You know what, we're big proponents of the community. We've been in it for a while. We just raised a new round of funding, and our first thing that we want to do with it is sponsor the community." And so it's like, great, I love hearing that.
Tony: So amazing and great way to just find your niche, find your model that works. And I'm curious about in terms of sustainability … sustainability of a community. You mentioned that the money goes back into the community. I know that the time and emotional labor of running a community is non-trivial. So I wonder how do you see that sustainability working? And I guess when you say you're reinvesting that money in the community, are you paying people with some of that money or is it all going to lots of different technology platforms? It's like a little bit more.
Jillian: All the tools.
Demetrios: Yeah. Well, there's one thing that is really cool for everybody that's listening that is a little bit of a hack. And if you reach out to companies, especially different companies that are starting, like Campfire or Gradual, like I said, and you just tell them, "Hey, we're a free community," and depending on how much revenue you generate from sponsorship or whatever, a lot of times people are at least going to give you a discount because you're a community and it's not like you're some company that is for profit. And so I've had a lot of success with that, just reaching out to different companies, especially in the beginning and being like, "Hey, you know what? We really want to use this or that, and we'd love to give you product feedback or whatever, if we can potentially start implementing your tool with our community."
That being said, though, to of the question, we pay people? Yes. This year is the first time that I'm actually bringing on a hire. Up until now, it's just been people in the community that have been working a lot, and then all of a sudden, I say, "Hey, you've been putting a lot of time into this community, why don't we start paying you?" And a lot of times it's hard to get someone to start taking money. I guess there is a little bit of things change when money gets involved, right? And so someone is doing it as a passion project, and they're just really stoked on the community because it's given them so much and they've been able to advance in their career or learn about different things or get a new job, whatever it is, then in a way, they want to give back.
And when somebody's given back for like three to six months, and they're really working hard on things, I have a really hard problem, or I have a hard time just letting that happen and letting them volunteer or do this for free, even though they say, "No, I don't want any money." So getting back to the question, we've been paying people who have volunteered. We have a few people that volunteer. Obviously machine learning engineers charge a very high rate, and we're not paying them anywhere near market rate, but they don't want to. In a way, it's like going back to that, what I just said. And now we just brought on someone to help me with what I call sponsorship success. It's like customer success, but sponsorship success, just because these days we've brought on a lot of sponsors, and I want to make sure that I deliver the most value to them as I can, and I don't want anything to slip through the cracks.
And so I want to do the under promise and over deliver, make sure that all of these sponsors, they come in as soon... When we're negotiating, I'm pretty standoffish, and this is the craziest part is because, again, going back to, there's a very small market in with the machine learning operations. And so I'm actively pushing away sponsors and telling them, "No, I don't take just anybody." And I'm actively pushing people away and try making it really hard for them to come and sponsor. But then as soon as they sponsor, it's like, you're family. All right, let's make sure that you get the most out of this because I want you the best experience, and you took that leap of faith. So I want to acknowledge that and make sure that it's worth it for you.
So we brought on sponsorship success. I also have a VA. Actually, I've had a VA since like last year. So that technically was somebody that yes, we are paying. And that changed my life too. Once you get to a certain level of doing stuff, I imagine you all know about it, but VAs are incredible.
Jillian: You know what, just me giving unsolicited advice because I like to do that. Especially now that I have a microphone platform, it's like, but it sounds like just listening to you, your community, like you said, like they get paid a lot in their day jobs and they don't want your money. So I would listen to them. And maybe, I mean, have you thought about doing like ambassador program or just some sort of like member recognition? If someone's just bang out amazing, send them some swag. I saw you have a shop, send them one of those cool shirts, or even just even like a card, or a badge in the community, which now that I'm thinking about it, Slack doesn't really do.
Demetrios: Yeah, Slack is hard with badges. And we've thought about one thing that I'm working with because it is, like I think one thing that is really cool with this space that we're in and machine learning and data and all of that, there's some cool stuff that we can just use the community to create cool stuff. And so AI-generated art is one thing that we've been looking into and creating different AI-generated NFTs for people that are in the community, and they're like, we want to give them a shout out and we want to tell them that we appreciate them. And so that's a project that some people in the community are working on.
And then another one is like, well, we've got all this data of people talking about all this cool stuff, we should create some kind of report, or we should run some machine learning algorithm on it, some natural language processing and figure out like what's the sentiment around different themes. So doing stuff that has to do with the community's day jobs and how they can actually put it into practice has been something we've thought about too, and we're actively working on.
Jillian: I wish I understood this on a deeper level, because that is so cool.
Demetrios: I don't either. I just pretend like I do.
Jillian: I didn't even know that was a thing, but now I'll go look and be like, "That's nice, I don't understand."
Tony: Demetrios, we have made it to the final bit.
Demetrios: Here we go.
Tony: We are going to take you through the rapid fire. Our rapid-fire has historically not been rapid. So this version is going to be rapid-fire now with more rapid. So keep it to maximum of one sentence. And if you could do it in one word or one phrase, that works too. Okay?
Demetrios: Oh, I like it.
Tony: All right. Okay, here we go. Starting the clock. There's no clock. What did you want to be when you grew up
Demetrios: A dentist.
Tony: Dentist. Amazing. How do you define community?
Demetrios: A group of people getting together that share the same goals and interact with each other.
Tony: Something on your bucket list that you have done.
Demetrios: Lived in a foreign country … living in a foreign country, living outside the U.S.
Tony: Something on your bucket list that you have yet to do.
Demetrios: Yeah. I don't know. I just don't have that crazy of a bucket list, I guess. It's not that I've had such a wildlife, but I have... I guess learn German. There we go, because I'm in Germany right now.
Tony: A book that you are loving.
Demetrios: I'm reading so many books right now. So I have a daughter, she's three, and How to Talk So That Kids Will Listen is a great one. But then more intellectual reading, I've been getting into The New World Order by Ray Dalio, which has been absolutely incredible.
Tony: I can't help, but ask follow up question. Do you find that the How to Talk So Kids Will Listen book applies to grownups equally?
Demetrios: I don't know. It's kind of low key manipulation techniques.
Jillian: That's parenting.
Tony: I mean, in truth, the kids are manipulating you. So it's only fair. All right. If you could live anywhere else, where would it be?
Demetrios: Shady Guaratingueta, Brazil.
Demetrios: Go Google that.
Tony: I'm dying to ask follow up questions. How do you want to be remembered?
Demetrios: As someone with a big heart who enjoyed giving and was a joy to be around?
Tony: It's beautiful.
Jillian: It's lovely.
Tony: How do you define community?
Demetrios: A group of people with a shared goal who are interacting with each other to further these goals.
Tony: Beautiful. Beautiful. Beautiful. Demetrios, how do we find you on the internet, and how do we find your community, and all the good things you do?
Demetrios: Yeah, the community's MLOps Community. I'm not sure how many people actually are going to want to go and get into the MLOps space, but we've got everything like YouTube and Twitter and LinkedIn. But I think I'm probably me, myself, I'm the most active on LinkedIn. I tried to write some tweets last night as like a thread and realized how bad I am at Twitter. So I'm not generally on Twitter, but yeah, LinkedIn is my space.
Tony: I can very much feel you on the, "I tried to write some tweets last night." That hit me right here. Demetrios, this has been amazing. Really appreciate your time. Keep up the amazing work.
Jillian: And candor.
Demetrios: Thank you, all. I appreciate you letting me come on here and just chat your ears off.
Tony: All right. Good stuff. That's our conversation with Demetrios Brinkmann of MLOps. I feel like I understand machine learning better on top of all the other stuff, which is just handy.
Jillian: I feel like it's just another thing to add to the pile of stuff I will never totally understand, but appreciate and makes my life better. So thank you to all you MLOps people out there.
Tony: It just goes to show that the value of a good niche and the people who do get it are really grateful to be in a room full of other people who also do get it, and taking it one step further, not just machine learning, but machine learning Ops, like that is an extraordinarily specific niche. And then being clever enough to recognize, kind of read the room and see that the opportunity there was in sponsorship. What did you think of that, Jill?
Jillian: Yeah. I mean, what a cool way to take a community model and create revenue in a way that isn't necessarily like a membership fee, although you could do that as well, you could do both, but also a sponsorship that is just super valuable. And the fact that those sponsors are active participants in the community, it's kind of a win-win, right?
Tony: Yeah, it really is a win. I hope that inspires everybody listening to just be thinking in terms of who's in the room, where's their value being created, or where could there be value being created, given the of people who are in that space. So the fact that this is one of the only places on the internet where you can consistently reach people of this exact niche is super valuable to people who are interested in that exact niche. And odds are that some of those potential sponsors or supporters are already numbers of your community. And so I'm left wondering, if I'm searching for new revenue streams for my community or a sustainable model, how can I be looking at who's in my community and what companies they work for, what those priorities are, and see where there are opportunities to partner up?
Jillian: Yeah. I mean, it's limitless, frankly. There's so many things you could do with this. I think where you have to be careful, and Demetrios is a great example of being careful, is it doesn't just turn into this spammy join my community and now I'm going to shove affiliate links down your throat and all these things and bye, bye, bye, because I'm just like, "Why am I here?" So that the fact that the integrity of the community comes first, but then there's these opportunities to engage with sponsor level people who are legitimately there to share and be a part of the community like chef's kiss. That is cool. That's very cool.
Tony: It can be a very tricky balancing act, but I'd like to hope that for any given community, for any given person running a community, there's a direction that will work, that will be sustainable, and it might not be the one that you originally came in thinking it would be. Maybe this model wouldn't work if he went in asking people to pay for a paid online membership community. And so by saying, "Okay, I'm not going to ask people for money and in so doing, I'm going to make it a really low barrier or to entry and rely on the fact that everybody here has enough shared common interest in the thing," that's going to drive things, and then we can find value that we can derive from that after the fact. But staying committed to that, making sure that you're not compromising the integrity of the community is so important, and it can be very tricky to get that right.
Jillian: Oh, for sure. It's a balancing act on the best of days. It's hard.
Tony: We'd love to hear what you think of Demetrios and all of his cool hacks. He mentioned some great websites that I'm going to be looking up as well. Let us know what you think of Demetrios and his community and what you're learning from it. And if you're, hey, ideally, get something from this, go find a new sponsor in your community or a new way of approaching monetization. We are @TeamSPI on Twitter, and we'd love to hear from you.
Jillian: Thanks for tuning in. We will see you next Tuesday.
Tony: This has been The Community Experience. For more information on this episode, including links and show notes, head over to SmartPassiveIncome.com/listen.
Jillian: Learn more about machine learning ops at Demetrios's community page. It is MLOps.community. You can also connect directly with Demetrios on LinkedIn. Just look at him up Demetrios Brinkmann on LinkedIn.
Tony: Our executive producer is Matt Garland, our series producers are David Grabowski and senior producer Sara Jane Hess, editing and sound designed by Duncan Brown, music by David Grabowski.
Jillian: See you next time.