• Ken Jee

Growing the Biggest ML Ops Community From Abroad (Demetrios Brinkmann) - KNN Ep. 88


Today I had the pleasure of interviewing Demetrios Brinkmann. Demetrios 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. He loves how communities tick and their inner workings so much so it has become one of his passions. When he is not conducting interviews you can find him making stone sackings with his daughter in the woods or playing the ukulele by the campfire. In this episode we learn about how Demetrios was able to grow an incredible community and some of the growing pains he faced when scaling, we also learn about how he avoids burnout and his love for ice baths.

 

Transcription:

[00:00:00] Demetrios: And a hundred percent tell people that are in transition, we could say, or they're looking for different jobs, or they are doing something and they want to go and jump into data or machine learning or whatever it is. You don't always have to be the person that is building the model or deploying the model to add value.

[00:00:30] Ken: This episode of Ken's Nearest Neighbors is powered by Z by HP. HP's high compute, workstation-grade line of products and solutions. Today, I had the pleasure of interviewing Demetrios Brinkmann. Demetrios is one of the main organizers of the ML Ops community, and currently resides in a small town outside of Frankfurt, Germany.

He's an avid traveler who taught English as a second language to see the world and learn about new cultures. Demetrios fell into the machine learning operations world, and since he's interviewed the leading names around ML Ops, 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 and ML.

He loves how communities tick and their inner work so much so that it has become one of his main passions. When he's not conducting interviews, you can find him making stone sackings with his daughter in the woods or playing ukulele by the campfire. In this episode, we learn about how Demetrios was able to grow an incredible community and some of the growing pains that he faces associated with that.

We also learn about how he avoids burnout and his love for ice baths. I had an incredible time speaking with Demetrios. I'm really glad I could share it with you. Demetrios, thank you so much for coming on the Ken's Nearest Neighbors Podcast. You've done such incredible things with the ML Ops community. I've only had like five or six people from the community say you have to talk to the Demetrios. You have to talk Demetrios and finally made it happen. I think Ben Rogan introduced us out. Mikiko also told me to reach out. Two people who I've had on the show already. So I'm like super, super excited to bring you on to talk everything, community building ML Ops, and your story.

[00:02:09] Demetrios: Dude, pleasure to be here. I'm so glad we get to do this.

[00:02:12] Ken: Me too. We've had a couple of conversations offline already, and I've just really enjoyed speaking to you. So that usually bodes well for the podcast listeners, also a joint conversation.

[00:02:27] Demetrios: I know there's a ton we want to get into.

[00:02:32] Ken: Absolutely. And hopefully we'll cover it all. So, if we don't though, we'll just do another one. So , the first thing I always ask people is how they first got interested in data or an even community building in your case. I'd love to hear a little story, a little background on that from you.

[00:02:51] Demetrios: Yeah. So, I fell into like ML and MLR really randomly.

I was teaching English in Spain right after I got out of college. And I was doing that for like eight, nine years while I was doing that. I was always trying to like create something or other, whether that was an immersion program on the countryside where we would bring a bunch of Spanish people to some cabin in the woods and we would only speak English.

So, hopefully, brought up their English levels or we were doing like pop-up art events and we were organizing these art events or craft beer was really big in the US but not in Spain yet. So, we were organizing craft beer events, but eventually I had my daughter and I realized that teaching English and all of this, like just me doing whatever floated my boat was not really that sustainable for like family life.

And there was a switch that turned in my head and so I wanted it. I was looking for something more stable and I ended up having a friend, tell me, Hey, I got this job as a sales person at this company. They're fully remote. You can work cause he was in the same city as I was. And he was like, you can work from your house.

No problem. Doing like SDR work. And for those who don't know what SDRs are, that's the people that annoy the shit out of you with your looking, they're trying to get you to buy stuff all the time. They're either cold calling you or sending you LinkedIn messages or sending you emails. And you're like, how did this guy get my email?

So that was me. And I'm sorry if I have disturbed anyone out there that is listening. I absolutely repent, but I was doing that for a while. And then that company wasn't that cool. I started looking for a job at another company still doing SDR work because at that point, that was all I had as my like track record.

I got this job at a company that was doing. They were really focusing on provenance for data science and specifically machine learning and making sure that it was like, kind of like a cube flow pipelines, but really trying to attack like, okay, if I change anything in my data, and then I put it into a model, how do I have this for data kind of thing, which is a little cliche these days. I know I've even said it a bunch, but in those days, in 2019, I had no idea that was cliche and it probably wasn't back then, long story short or long story longer. I ended up getting that job, working for as an SDR for them but the company went out of business, right when COVID hit.

And about a month before COVID hit, we started doing meetups around ML Ops, and now our CEO of the company, that company was called .science. And then, the second meetup everybody was like, Oh, I'm too busy to do anything. And so instead of doing it like a meetup, I had always wanted to run a podcast and I just was asking questions.

And so I started organizing guests. And then that happened as a podcast that like became the podcast, which is now the ML Ops community podcast. And then the company that I was working for went out of business and I had this choice, like, Should I continue with the podcast and the community? And it started out as a meetup, but then really early on, we could see that there was a lot of interest in ML Ops.

And so we started being like, should I continue with this? Or should I try and go and get a job somewhere else? And I applied for like a million jobs. Those who are listening probably remember in April 2020, nobody was hiring. Everybody thought the world was coming to an end. And so yeah, that's how I got into data. That's like the longish version .

[00:06:49] Ken: I love that story. I mean, to me it seems like you've always been passionate about community building, right. And you've been able to evolve that over time. You're able to find a niche where people really valued the community building aspect and you know, it didn't take you having these crazy.

Like ML Ops or data science, technical skills to build a community. It took actually listening to people and storytelling and this like almost completely different skillset than what makes someone traditionally successful in like a purely technical role to make that happen. And I think that that's a perfect highlight of there's like this beautiful symbiosis and or there can be this beautiful symbiosis of people who who like traditionally didn't come from this like super technical role that can melt into this world and add so much value and so much you know, like so much glue that keeps everything tight together, or helps everyone communicate, or facilitates things. I mean, obviously you've picked up quite an understanding of ML Ops at this point, I would imagine. But, you know, you were able to create value without necessarily having this like massive understanding of ML Ops to begin with.

And I really wish a lot of people would realize that about this domain, right? You don't have to be like, I am by no means the most technical person, but I still make videos on the internet that are useful to people that are, you know, hopefully gonna surpass me one day or whatever it might be. And this idea of community is something I want to stick on a little bit.

So, you know, like how did this community grow? You know, since you started, you decided to stick with it. What are some of the things that you did that made it expand, or that really were linchpins where you're like, Hey, I think this is going to become something as great as it is.

[00:08:45] Demetrios: Yeah, there's a few pieces of that, but just before we jump into that, I want to address that point that you were talking about before, because I a hundred percent tell people that are in transition, we could say, are they're looking for different jobs or they are doing something and they want to go and jump into data or machine learning or whatever it is.

You don't always have to be the person that is building the model or deploying the model to add value and all these, like if I look at all the ML Ops or ML tooling landscape, there's so many companies right now that are just getting money poured into them by venture capitalists. And they all need like marketing teams that kind of understand what's going on, but are also really good at marketing.

They need sales teams, they need product managers, technical writers, recruiters, people that there's so many different places that you can fit in. So if you do not enjoy writing code, or you're not like the biggest fan of Python, that doesn't mean you need to totally exclude yourself. I mean, yeah, designers, you need that, but I try and tell people this when they're in that position, which is, if you are looking for a new job, like the best way to do it is just put yourself out there. Whether that be writing a blog or starting a YouTube channel, or just going to, or presenting at different events, trying to get yourself out there and build that presence.

Even just like writing on LinkedIn and doing all of that is super helpful. You know, just as much as I do, but that's a side note getting back to the community part, but let's talk about that. Like, how did I know that things were happening? What were some big pieces or big signals that this is going to be something bigger than just like a couple hundred people in a Slack workspace.

It was when, so first of all, it was growing really fast. But in the beginning I was reaching out to a ton of people on LinkedIn and telling them, Hey, I just spent like 40 days inside, which was not a lie, but I had my sales background in, which was just cold reaching out to people. And so I just started cold reaching out to people and telling them about the community and saying like, I just spent 40 days inside and I wanted you to do something useful with my life.

So I started this online community. Join us if you want, I can give you more details. And so I was doing that. And then I stopped doing that at some point because the job hunt was real and I was just trying to get a job. And so I didn't have time to like cold reach out to hundreds of people on LinkedIn.

But what I saw was that people were still coming into the community, even without me having to go find them on LinkedIn. So that's the first thing that I knew that I was like, Whoa, wait a minute. There's something here. It's like organic growth instead of me having to force it. So then the other part was I would get people that were really interested in the topic and really interested in the community itself and just volunteering right away.

Like one of the main organizers of the community is David Aponte and another one's Vishnu. And both of those guys were in the community, like from the very beginning. And they raised their hands and said like, Hey, if you ever need any help, just let me know. And so having those guys, having just people saying that and then being able to go and look at who is more technical than I was and asking them, doing things that don't scale. This is almost cliche by now, but one on having like one-on-one conversations with people that I knew were really like battle-hardened veterans in the ML Ops space. And I would ask them, Hey, if you know, if you see any questions that haven't been answered, do you think you can go and answer them?

Or if there's any articles that you like, can you just share them with the community? Just so there was movement and it wasn't me like taking every post I saw on LinkedIn about ML Ops and then resharing it in the community.

[00:13:10] Ken: I love the idea of just that interaction, how do you foster that. That's something I've, you know, in my own communities that I wish I could do a better job of is just celebrating the people that do interact and giving them at least some feeling of ownership and those types of things.

I think from the other side of that consumer perspective, people also don't realize how much that can benefit them. I mean, there's a couple people in some of my Discord servers, in my communities who I've like given paid work to. I'm in the process of bringing someone on my team who is actually like a high school student, right. That's in another country and it's like, Hey, they're doing incredible work here. I've been able to, like, they've been doing it for free for almost a year, a year and a half. You know, they've proven to me that they can do great stuff. Why, why not get them involved? Why not scale in that way. And you know, you were talking about doing non-scalable things.

I think communicating with people is actually one of the, or finding people that are like truly game changers is actually one of the most scalable things you can do, right. Is that if they can do everything that I can do and that I could focus my efforts on the things that I'm strongest in. I mean, that's how you're able to grow so effectively.

Can you talk a little bit about, I guess, how the podcast fits into that and like the story of that, and you know, like the scope and some of the most interesting moments there.

[00:14:38] Demetrios: Yeah. And it's exactly that, like, going back to that point, you were making with people that are active in the community, whenever someone reaches out about anything, who do you think are the first people that I think about, right. The ones that I have the closest interaction with, or the ones that are the most vocal or the ones that are active in the community. So if somebody reaching out about expedites, I'm going to introduce them to the people in the community that I feel are the most adequate. And the most adequate people are the ones that are usually like raising their hands or being active in Slack or wherever it may be, or that I've interviewed or whatever.

And so coming along to the podcast, we started doing not just the meetups, but like podcasts because David Aponte, when he, he like reached out and was like, Hey, I want to help. And what can I do with the community? I said, you know, it'd be awesome if we could just, I feel like I'm just barely scraping the surface when I interview people.

If I can have somebody that either also interviews or is interviewing like a co-host with me and is able to go deeper than that might be like better for the all-round health of the podcast. And. I also was using it. This is just like another side note, the beginning of the podcast and the beginning of the meetup when I was without a job, I went and I emailed and I like, reached out to every single. And when I was interviewing them or after the fact, I had created a relationship with them and then I would try and ask them for a job. So it just so happens. Like by the time I ended up like interviewing people, I was like, Ah, well, I don't really need a job anymore. And so that didn't matter as much, but that was the idea there.

It was like, I'm going to interview these different technical founders or whatever, and I'm going to use it as an excuse to get, to build rapport with them and to get in contact with them. Because if it's just me as somebody sending a resume, it's a lot harder to create that connection. Then if it's me with my podcast, Hey, you want to come on my podcast.

Cool. And then you have that access to people that you don't normally have access to. So that was the start of it. Since then, it's been able to grow into something where it's like, again, I get to D and he wrote the ML Test Score. He wrote the high interest credit card debt. He's writing a book right now called like reliable machine learning.

Some people have dubbed him the godfather of ML Ops, and it may or may not have been me that said that somebody called them like the grandfather of ML Ops, but I think godfather is cooler. And so I got to interview him, which is like a dream come true because it feels like he doesn't ever do media appearances, but because there's other people in the community that I've interviewed, and they really liked when I interviewed him, they were nice enough to pass on an intro. And then he was, he came on the podcast and we got to like learn from this just legend. You know? So that's in my eyes. The podcast is awesome for those reasons.

[00:17:57] Ken: This episode of Ken's Nearest Neighbors is brought to you by Unlocked by Z by HP. Honestly, this is by far the coolest project that I've ever worked on. Z by HP has created a movie specifically for data scientists. Not only is the movie entertaining, but I personally created 4 data science challenges that are integrated into the plot. You get to work on the same data that the data scientists in the film are analyzing. Nothing like this has ever been done before, and even better. If you submit your work to the beginner friendly challenges, you get entered to win one of the new Z Book Studio laptops, the exact same one that I use. Check it out now at hp.com/unlocked. Again, that is hp.com/unlocked.

So to me, I agree with you a hundred percent on the podcast off. There's something really beautiful. And I don't know, at least for me, I don't mind saying it. It's like a little bit self-serving that I predict that I produced the podcast, right. I got to talk to incredible people. They recommend me to talk to other incredible people and I make new friends, but I also build my network in a very unique way where it's such a deeper relationship to everyone that I bring on.

Then if I was to just reach out on LinkedIn or to make a connection request or whatever it is, right. You you're sitting, you're getting to know someone for an hour rather than just sort of, kind of like loosely saying, Oh, I have like have some connection to this person. And we've never really talked that much or anything like that.

And the opportunities that come from the podcast ,,,, incredible. I mean, you talked about, Hey, this is a great way to find a job or get work connections or whatever it is. And I've seen a lot of college students have, you know, produced podcasts or produced some form of interview style content that have led them to unbelievable success.

Because as you had mentioned, I mean, this might be a really powerful sales thing. If you're, if you have a reason to talk to someone that isn't completely self-serving right. If you're like, Hey, I can help you tell your story. I can give you this audience. I can, if you care about something, we can help you put this into the world.

Everyone wins. There's no like, Oh, I'm getting something from you. It's no, we're both benefiting. And I really like those types of relationships, right? I mean, I look at YouTube like that, right? Where I'm putting value into the world, you know, frankly, I also make a decent living from the content that I produce there.

And I just love seeing other businesses where it's a win-win and and people are able to create systems like that. Completely, you know, you have talked a little bit about, you know, bringing people on and essentially like the ML Ops community is self-sustaining. How did you manage delegation?

How did you manage working really hard on this and not burn it out.

[00:20:43] Demetrios: Yeah. So that's something that I have to constantly be coming back to and like reflecting on because in the beginning it was completely a joke and because it was so new and it was basically, I would make fun of myself. Like whenever I would go into a podcast, I would be like, Yeah, I got to go build my personal brand.

And because I was so far outside of the ML and ML Ops universe, I felt like there was no way that anything I ever said people would ever like take seriously. And so then from there I went and I was like, Okay, cool. How can I make it the most fun, possible I was making cheese, so I remember we had one guest on and I was like, This episode is sponsored by cheese and look, you know, you have Swiss cheese there, this cheese, and it was just making up sponsors like, and having jokes about it.

And then with David like the first, first couple of episodes, we were just. Riff on each other's shirts and try and wear like the most outrageous shirt possible. And then like the guests wouldn't know about that. Yeah. I was going to mention your shirt's pretty sweet. And so it was bringing back old memories.

Yeah. And so we were doing that then later, like the first sponsorship money we got, I went on to Amazon and I bought the most ridiculous sweater that I could find. And I sent it to David and I made him wear it for the first for the next like podcast appearance. And so we were just doing like all kinds of fun stuff like that.

Eventually like you do enough podcasts and then the energy kind of gets lost. It becomes more work. And so I always have to re-correct and come back to that center and like realign, okay, what am I doing this for? Because I'm not the community itself is self-sustaining. I'm not like making my living out of this.

I'm just doing this out of pure joy and passion. And so what. Like w how do I want to, what kind of self do I want to bring to this? So I want them to just be like, Mm, getting through this, because I scheduled way too many podcasts, or I over committed to way too many like sponsors. And now I don't want to do this sponsored talk or whatever, or do I want to like, bring that fun into it and keep that fun and interesting vibe.

And so it's a constant coming back to like, Oh yeah, I'm doing this because it's fun. And I enjoy it. Not because I have to and it's work. And so keeping that in mind has been crucial for me. And I think that it gets reflected in the community, right? Like it gets reflected in the way that people treat each other in the community, because we all are in a way exploring this new space together.

And nobody is taking themselves too seriously. And when anybody does that, they kind of, I kind of called them out right away. Like somebody, you know, like if they say like, Oh, and they come in and introduce themselves, and they're like, I'm an expert I've been working in the field for 10 years. It's like, Well, okay, cool.

But you know, I've been around like five years. Yeah. And it's like, even these people that I've interviewed that are what I would consider the cream of the crop, they don't even call themselves experts. So you might want to check your tone, but anyway, the burnout is, is one of those things where if you're not careful, and if you're not very conscious about how you're doing it, then it's very easy to creep up on you.

And so whenever I recognize, like, Oh, I'm not being as playful anymore. I just try and be ridiculous. And over the top and make jokes and remember that. Yeah, there is part of this, which is people are tuning in for the learnings, but there's another part of it where it's like the occasional laugh. Isn't gonna hurt anybody.

[00:24:54] Ken: I love that you're able to keep, keep it sort of light in the communities. Right. That's something I see. So on like Reddit, for example, I think people take themselves away too seriously and that you can, because it's anonymous, right. Where it's like, Hey, like you say something and you have like 20 people coming in to like, actually it's like this actually it's like this.

And it's like, no, like, you know, this isn't, you're not trying to go to places for debate. You're trying to go there to learn and to grow. And you can do that in a fun way. I'll also say, at least for me, I've learned a really valuable lesson in content creation or putting stuff out there. Is that again?

At least for me. I am have to find enjoyment in it. I have to like want to make the content for it to be of any quality. I mean, there are plenty of things that people want me to make or people that want me to interview them. And I'm just like, I'm not excited about that conversation or right. I'm not excited about making that piece of content.

And it's nice that you have the ability to like say no and be like, Hey, I don't think that this would make the best possible thing for this medium or, you know, you're not going to be like, no, I don't want to interview. You're boring, but you do have the power to be able to like make decisions around that when it isn't just like work, you know?

And I think that there's also a really cool benefit of like, looking at these things as not the only thing in your life. But there is, it is like one thing that you derive joy from, and then eventually you can also make income from I talk repeatedly about how I probably won't ever leave my consulting work because that provides necessary balance for me to create the best content that I can.

Yes, I could spend more time creating content if I wasn't doing that, but I would also have a different incentive structure that would be paying my bills. There would be different decisions that I would make. And I can't safely say that I would make the same decisions that I make now that I think are in line with the best things for my community.

[00:27:03] Demetrios: That's super important. Like, you're not so attached to it. And you're, you don't have to have that stress about like, where's I have to get sponsors or I have to create something because that's where I get my next meal. And if you don't produce it, you don't have that. I think if you can detach the two, if you're lucky enough to be able to detach the two, then it's great.

But also there's the other side where it's like, well, you could be missed a beast and you just pour everything into it because it makes you a ton of money. And so who knows?

[00:27:35] Ken: Yeah. I mean, I don't think that there's a right answer. I know that for me, that's like a difficult thing that I would have to, like, it'd be a big time introspection look in the mirror conversation.

Like how do I do this without compromising. What I think is appropriate for my communities and things along those lines. I am very interested in sort of your living situation that you're in Germany right now. You spend a lot of time in Spain. If I recall you're from somewhere in the U.S. But I don't remember exactly where, how, how did all of this love for travel start and you know, what are the implications for the pandemic and also like community building?

[00:28:17] Demetrios: Yeah. I got the travel bug real early, man. Like my mom took me, so my mom's whole side of the family's from Greece. And she took me to Greece on and off occasionally when we were growing up and when I was younger, I didn't like it as much. But then as soon as I like hit 16, it was like, whoa, creases. Cool.

And so from there, yeah, exactly. And so then I started traveling. I really enjoyed being outside of the U.S. just because it wasn't English being spoken everywhere. It wasn't my culture that I knew and grew up with. And so I had to be on my toes a little bit more. And then from there, I remember I went to, when I was in Greece, one summer, I went on this like on a sailboat to Turkey and I experienced Turkey and that felt like a world apart from anything that I had ever encountered.

And then from there I went to a I went and I was going to university in Spain for a little bit, doing like a semester abroad. And in Spain, I went to Morocco on like one of the many holidays that they have in Spain. And so I went to Morocco and then that felt like even more, like further away from even Turkey.

And so I caught that bug and I wanted more of that. And when I finished university, I went to India and that was just like, I dialed it up to 11 along the scale. It felt like nothing that I had ever experienced before. And so I stayed there and I was volunteering. I was doing like volunteering for an NGO for like four months.

And I just had the time of my life. Like I loved everything about being in India and this. Like forcing function that you had to be outside of your comfort zone every time you went out of the house. Like the only place that was comfortable was in the house when I was playing guitar that I bought on the street.

And so I was just pushed outside of my comfort zone. And then I came back to the U.S. and it was like I can't, I'm going to keep traveling. I want to see if I can try and do this, like, and live outside of the U.S. So that's why I went the easiest way to do that at the time, like in 2010 was to get a job teaching English in Spain, because I had settled on Spain because I was, I spoke Spanish and I wanted to, I really enjoyed the Spanish culture from when I was there doing study abroad.

And so I was like, all right, I'm going to go to Spain. And I had met a girl in India that was coincidentally living in Spain that may have had something to do with it too. And so I settled on Spain and I stayed there for like the last nine years we came during the pandemic. So everything kind of happened at once.

Like the pandemic, then losing my job and then us deciding to move to Germany all at the same time. And so there was a time when I was and starting the community, I was starting the community. I didn't have a job. And we were planning on moving to Germany and the pandemic was in full swing. And so we came here just because it was like, there's a meditation center that is really close to here.

And we had been coming like, we were familiar with it. And we had been coming to it quite a bit while living in Spain. But then we had a nice group of friends built up around here and we said, man, with the pandemic, living in the city is not as much fun as it used to be. We want to get out right now.

We're living on the countryside. There's nobody in like around. And that's a blessing because I'm outside of that phase of my life, where I need people around. I think when I was younger, I probably needed a lot more people around and I wanted to go out and do things. Now I'm much more apt at going into the forest or going on a long run and not seeing anyone and just seeing like wild hogs and running from them or foxes and deer and all that.

So, so that's what I've been up to now. And yeah, when we came since then, I mean, haven't looked back, man. It's been great.

[00:32:53] Ken: That's incredible. I mean, it's something I'm interested in. I do want to touch on a couple more things about the community, but while we're here, I'm really interested in kind of how meditation and you also had mentioned getting out of your comfort zone was a big part of travel. Yeah. How, how are those two things related if they are, and you know, what have they meant to you as, as you've developed, it doesn't have to relate to the community. It can be, but those are things that have become very important in my life. And I'm always interested to hear what they, what they mean to other people.

[00:33:27] Demetrios: Yeah. I mean, when you look at like traveling and especially for a while, I was traveling alone and in India, I went by myself and you have to be your own best friend because you're like your only companion. And that makes you have to like, come to know yourself a little bit better and know what kind of thoughts are coming up and know what is happening inside of you when you see different things that you may or may not.

Appreciate, or you may or may not be familiar with, or they just are testing your limits. And so, because of, I mean, I was meditating before I went to India. I meditate. I started meditating around 19 just because I had traveled in Greece and I was traveling alone then too. And then I got back from Greece and it was like, I need to know, I need to get more familiar with myself.

And it just became this journey inwards and the external journeys are great, but the internal journey is the one that without sounding like too psychedelic it's really the one that is the best journey that you can go on. And so I became, I try. I remember setting this goal for myself. Like I want to be my own best friend.

I want to enjoy my own company as opposed to having my mind and my loneliness be something that I was afraid of. And I was always distracting myself from, so I started meditating bunch and then traveling alone. And it really, at the end of the day, it just helps, you know who you are and it helps you become more comfortable with who you are.

And so in that sense, like in the community and building the community. I'm much more comfortable with saying that I'm not technical or saying like different things that I feel or making fun of myself. I mean, just before we hit record, you saw the video, I put it out. It would be hard to make it like that.

That's for everyone who wants to see a funny song of me singing out a tune that is the video I put out. And so I do all of those things, like, just because I'm not so concerned about who I am, that other people feel like that, or like what other people think of me. I'm more concerned about my own image and my own ability to be with myself.

So that's what meditation has ultimately brought me. I would say. I mean, besides all of the mystical experiences that one has while they meditate, but the best thing that I've found. Yeah. Is just to be able to be with myself and be at peace. And while I'm with myself, right. Instead of being anxious or looking for a way to distract myself.

[00:36:37] Ken: Yeah. I mean, I think that that's really beautiful and I think it's also something that a lot of people, especially earlier in their careers, which I see a lot that they're struggling with, right. They, they hear data science is cool and sexy and you make a lot of money or machine learning and they hear these terms and they don't evaluate their look internally about if this is something that they are interested in, right.

Like that they would generally be a good fit for. And it's frustrating to me because you know, people are wasting their own time, a lot of the times, and that leads. And I also think that there's this really unique connection between understanding yourself and exploring new things, right? Like there are two sides of the same coin.

So when you're, when you're looking internally when you're meditating, when you're doing these introspective practices, you create this image of yourself in your mind, right? You say, this is who I am. These are the things I believe. And when you push yourself out of your comfort zone, you're testing that in the real world and you're getting more data.

And you're, I have a very weird data centric approach to to mindfulness. But you know, to me that, that, that is, that's like a cool feedback loop of all right, who am I? Is that really who I am? Do I really like these things reevaluate this like beautiful cycle of like understanding and testing and like one of my favorite stories, right?

It's there's this monkey goes on this three-year meditation retreat and he thinks he's found himself. He thinks he's won. And he exits the cave and he goes down to the street and a farmer cuts them off. And he feels a little bit of anger and then he goes back up for another three-year retreat because he, he like, wasn't where he needed to be, and like, that's like such a cool it's such a cool experience that someone can have the personal journey and the introspection is something I think everyone can benefit from. And it's also interesting to look at it as a system or as a lifestyle or as a construct rather than just like, Hey, that's this thing that these people do in the woods and, you know, they're all hippies and whatever.

It's like, no, this is, you know, this is a beautiful way to like forward everything in your life. And you know, I think that's really cool.

[00:39:02] Demetrios: And also back to your story, That's why I feel like I met its state is to be able to go into the world and bring the same presence into the world with me as when I have sitting down with my eyes closed, it's not just when I sit down with my eyes closed, I'm like at peace, I want to be at peace everywhere I go.

And so the more that I practice that I look at it like a muscle that I'm exercising. And so when times get tough and I have a taxi driver cut me off or something, doesn't go my way. I can exercise that muscle more. And it's already built up, like I've built that strength. So things don't affect me as much.

[00:39:48] Ken: Yeah. One, you know, there is always. Kind of drawbacks. So there's always, I think, negative things that happen to everyone. And it's a lot about how we frame them and how resilient we are that makes us successful and anything that we do. And so, you know, I'm interested a little bit in some of the bottlenecks that you've had with your own community, you know, what were some of the challenges that you faced?

I know that there are always growing pains. Right. And how did you resolve those with this zaniness that, that you have?

[00:40:21] Demetrios: Oh, my God vendors, all, that's probably like the biggest pain in my ass ever. And I'm always, I'm always yelling at the vendors because they just come in and they. Equate them to either like three-year-olds or dogs that are looking for fire hydrants to peak on because they come in and they spam the whole Slack workspace about their new tool or how we do this or that, and why you should check them out.

And they don't even introduce themselves. They don't even add any value. They just go right for the jugular, which is probably karma because I totally understand I was in sales. I know that people have to hit their numbers and they see a community with 8,000 highly targeted people in their niche. And they're like, Oh my God, I'm going to actually channel this whole place.

So had to learn how to turn off at channel really quick. Then I had to learn how to ban people from the community. And we had to iterate, constantly iterate on the Like the docs or the code of conduct is what we call it. So iterating on the code of conduct, trying to figure out, like, how are we going?

What is the protocol for vendors to come in here? Because a lot of vendors have a lot of wisdom and they have knowledge and they have the ability to share that knowledge. But if a vendor just comes in here and says, try my tool, that'll fix your problem. Then you get 10 other vendors saying that same thing.

It is not really a community. It's just like a spam fest. And I had to work really hard at that and create channels with all of the vendors and tell them, look, everybody, this is not cool. Please do not do this. And then it would happen again. And then it would be like, have to change my tone a little bit, and then it would happen again.

And he'd be like, all right, look, you're getting. See you later. And so no more, Mr. Nice guy. Yeah, I have to, we were just talking about how I'm at peace and all this. I was at peace while I was banding them. I'm not going to lie and be like, no, I wasn't affected sometimes like things that probably affect me the most are when I'm like, Oh, I'm going to take this day off and just hang out with my family.

And then when I come back to Slack, after taking one day off of Slack, it's just full of spam from a certain vendor, and that frustrates me the most. It's like, Ah, the one day I take off and I'm not around the computer at all. And I hang out with like my daughter and now I've got this to clean up. So they usually get like more of a wrath than just if you were to spam.

And I was online already and I caught it after like a minute. But yeah, that's probably been the biggest thing. And then you add on top of that, the extra layer of complexity. So it's like a complexity sandwich. Okay. Now we're going to take money from vendors because we have sponsors. So what happens when a vendor who is a sponsor feels like they are afforded extra rights or things that they should be able to do that others shouldn't just because they're now a sponsor.

And so that's a lot of fun and figuring out how to walk that line and like really tip toe around putting my foot down with the sponsors, but making sure that it's very clear when you sign up as a sponsor, there's no nothing that will lead you to believe that you now can spam the community. And so that's probably been like the biggest learning thus far. Just how to interact with the vendors.

[00:43:58] Ken: Well, I got to I know you're a Slack loyalist, but so all my communities are on Discord and I have a couple, you know, like one of the incredible people in my communities they've made some bots that are like unbelievable at detecting like vendor speak or spam.

And they like quarantine it and let us go through it. And I'm a huge fan of like technological solutions to keeping your head your head clear. Like I just started using...

[00:44:35] Demetrios: That all like the vendor's spam content went down drastically when I started using this app called Campfire.

That as soon as you sign up for the ML Ops community Slack, you get a, like an onboarding message from me. And it's really a bot, but it's me. And I'm walking you through this guide and it says, oh, if you are, what do you consider yourself? Like practitioner just starting or a vendor. And if they click it, doesn't say vendor, it says working at a tooling company.

And so if they click that, then automatically it goes, Okay, read our code of conduct. Please do not spam. Please check this out. And it shows them all these things, because it's a little bit of like, choose your own adventure in that if you click something different than you get different messages. And so there's that decision tree that you can go down.

And so that cut down a ton on the spam, just because people were. People that were working at ML tooling companies became educated on what they could and couldn't do. And what was all right.

[00:45:36] Ken: I love that. I mean, that's something that is so necessary and any community that has, that has grown. So the community I think, is around what it's like, it's 8,000-ish people now.

Right? Wow. How's like, what's the experience of going from like a hundred to like 2000 or, you know, even to 8,000, what are some of the different things you have to be conscious of aside from the vendors? Obviously.

[00:46:04] Demetrios: So, from like 200 to a thousand, I didn't really notice much. Just because again, like I was trying to find a job I was trying to move and I was and then at that same time when we were in pre 1000, I got a job building, another community that did on Kubernetes community. And so I was just doing both of the communities and I can't really remember what was happening, but I do remember that there was this sweet spot between like, it was in like November last year when we hit like a thousand and it was between 1,000 and 2,000 people.

And it felt like everybody that was in Slack knew each other. And there was a lot of momentum around it and people weren't afraid to get into conversations. And it, it was like just enough people there to talking and asking questions and helping each other. And then as it started scaling more and more, then I started to hear feedback from people being like, Dude, this is too busy. Like I can't keep up. There's so much going on here. And this is one of my, I know your Discord guy. And this is one of my favorite things about Slack is that you've got threads and you can have. These conversations tied to a thread. So really like making sure that people follow the rule of like sticking to a thread and not just sprawl and then trying to address different needs and break conversations into topics or into channels.

So that if someone's really interested in a channel, they can subscribe to that channel almost like subreddits in a way. But I definitely think there was a a few people that they are more in the small community phase. They liked the small community feeling, as opposed to now where it's just like gigantic community.

And I don't know who is answering. It's awesome that there's so many people answering, and I love that, but sometimes you don't know the person that's answering, they have great information. They have great wisdom and it's very kind of them to share, but it's not like that when it was like between 1000 and 2000 where we knew everyone.

Now I would argue that it's higher quality because there are more people and whatever your niche is or whatever your problem is that you get into there's somebody else that's been through that problem already. And then they generally answer. So those are the things that I've noticed as, as we scale.

The other thing that I'm really conscientious about right now is like, how do we make sure that the ML Ops community is not just a Slack community? So we've created this. I'm super proud of, but we created like community doc. Or a community orientation page and we can leave a link to that below, too, but it's just everything that we've got going on in the community, because we have a reading club, we have engineering labs, which is like a hands-on it's like a Kaggle for ML Ops.

We've got round table discussions. We've got the meetups, we've got the podcast, we've got the newsletter, we've got the blogs, we've got the tooling comparison page. There's just so much going on that somebody who first comes into the Slack is like, Okay, cool. There's the Slack, and they go to a channel, but then later they are in the Slack for two months. And then it's like, they see the meetup posts like, Hey, there's a meetup going on right now. Come if you want. They're like, wait, what? There's meetups. I didn't realize we did meetups. And you're like, yeah. Then there's a podcast. And then there's like a newsletter and all that.

So we put that all in one place. We also have some cool, like open-source projects that are basically being incubated inside the community. And so we put that on the docs and like, how do you start a new initiative if you want to start that, like, who's the lead of each, like the reading club, the aid or the MLF Stax lead, whatever it is.

And so that's been really useful to helping us scale and we're starting to also get like local champions. So people in different areas, if they want to do something in real life, they get like all of the power and weight and the whole like ML Ops community machine, behind them to do something in in real life, in their local area.

[00:50:31] Ken: That's incredible. I think one of the ways that you can bridge that gap of being, not just a Slack channel, just like you said, with the docs, but the more in-person stuff, like that's one of the things in my mind that fixes some of those growing pains associated with like, Oh, I don't know who any of these people are.

That's a whole different ball game when you meet people face to face. So like, Oh, I see that person in here. And it's like, there's that whole level of, I don't know, just like how our mind works. Like, there's this connection. I want to be more engaged. I know these people now. I really, I don't know. I just think that there's something so impressive and like fascinating about that.

You have developed obviously a lot of things. So you said the newsletter, the podcast, all these different things, and you're always trying to improve the community. You'd mentioned something about iteration before, and I'm interested in how you iterate on new ideas or how you put them out there or how you build these things.

[00:51:25] Demetrios: So I'm not like the most advanced pro at this, but I think I've, I've practiced some of the things that I've learned about. And one of them, at least for anything, community related, if we can get it out as fast as possible, then I'm going to do that. Like, I'm just going to try it and see what happens as opposed to creating a whole plan around it.

And then we're going to do this and then we're going to do that and like have a notion doc and socialized in it with other people and see what they think. And then we execute on this. I'm more of, I think I've heard, it explained like where, where. Where Google is that like, get it out and iterate. And then you have Apple is like throw one dart and hit a bullseye.

And so I'm more of the Google style where I just want to get it out. I want to iterate if it works great. If it doesn't, we reassess in like a month or three months, however long it takes some examples of this are we try to do this thing called office hours. And that was where we were just opening up a session and people could come in and talk about what they're working on that did not work.

But what happened was we found out it didn't work because I just opened it up. I said, Oh, people are asking for this, let's start a channel. Let's create like the meeting link. And then let's have it be a two week recurring event. And if people go great, if they don't, then we'll find out soon enough. So just trying to get things out, trying to see how it goes as fast as I can.

That's one thing that I really am a proponent of, like, especially with the community, because the community will tell you. If it's going to work or not like though it'll live or it'll die and it's going to be pretty obvious the next thing that I'm working on now. And I was talking with Joe Reese, who you may know, but he...

I figured he's awesome guy. And I was telling him, like I put on LinkedIn and the other day I was like, ah, it'd be cool. I'm thinking about writing a book around like interviewing different machine learning engineers and what they do all day at work. Like kind of like a machine learning engineers at workbook.

And is there any interest in this just to put out feelers and see if people were interested in that? And I got back a bunch of good feedback and then Joe was talking to me about it and he was. And I told him, I was like, yeah, right now I'm at this point where I'm deciding, do I want to put out that book?

Like, do I want to go and interview different machine learning engineers and learn about what they do and how they do it and see their different use cases or the other way that I can spend my energy and my time. And my focus is I want to create like a community course or like how to build a tech community.

Kind of just sharing all my wisdom into that. And Joe was like, well, why don't you just write a blog post first or create a short YouTube video and see if anything sticks and see which one gets more traction and then go after that because there's no use in creating a whole course. Even if you do iterate on it quickly, if nobody actually cares about that course, or your reach is fairly limited, then it's going to be a total waste of time.

And the same goes with this book, like write a blog post first as a one-time thing. See if people like it, see if it generates any interest. And then if it does, you can think about like creating and adapting it into more of a like full-fledged book. So that was huge advice from him. And so it's like, what is the, how can I start small on this?

How can I do something that's really simple and just get it out and see if I can get that feedback loop going as fast as possible. And whether that is like creating a channel in Slack for the community, or that is like putting out a blog post around, I'm thinking about doing this. And so let's try and do this first as a little bite sized piece and grow it later.

[00:55:28] Ken: I love that. And so it seems like a lot of these things are a little bit learnings from the engineering mindset. Right? I think that it's really cool that you're able to integrate that into a community or into non-engineering things. I will also say if you ever want a co-teacher on a tech building community building I am more than happy. That sounds really cool.

[00:55:57] Demetrios: Cool. We may have to do it, but I'll tell you how the blog post goes.

[00:56:02] Ken: Oh yeah. I'm happy to share when you do put out that, that blog post. So kind of the last thing that I want to touch on is we spoken we're both fans of ice baths and it does tie into the. Pushing yourself out of your comfort zone and the meditation and the mindfulness thing.

But I'm just interested to hear more about that. When did you start doing that? Definitely a fun little seminar related thing.

[00:56:32] Demetrios: Yeah. And for those who do I spasm they know for those who don't do ice baths, they think we're totally crazy. Because for most people, I think they think about going into freezing cold water and they're just like, that does not sound fun.

I don't know why you would ever do that, but it's that after effect, right? Like the, I was listening to the Huberman lab podcast. I don't know if you ever listen to that one, but yeah, you incredible podcasts. And he talks about dopamine and what happens to your body when you go in and you have this cold therapy, we could call it, I guess.

Cold exposure is what the correct term. And so what happens to your body after you go through this cold exposure? As far as your brain chemistry. Instead of having like a spike in dopamine that generally happens if you eat chocolate or if you like a snort, some cocaine or whatever, you get really high dopamine levels and then it just crashes, and then you feel like crap, but with ice bags, you get a very gradual increase and it can increase sometimes even more than what you would even get from taking a line of cocaine. And so that, isn't why I started it, but I found that out and it just reaffirmed it. And so for me, the reason I love I spazz right now, I only do them in the winter.

And it's kind of crazy, dude, because I think I mentioned this too, but I cannot. I have a harder time motivating myself to do an ice bath. If it's not super gnarly, like if it's snowing out and there's ice, that's in the ice bath and I have to like break the ice and then go into it. Or it's like, I have, I'm going to go and swim in a lake that has ice all over it.

And you know, there's snow everywhere and it's negative five or whatever, then I'm all about it. I'm really like, I don't have a hard time going and doing that. But if it's like, so I'm in Celsius. So I guess like if it's like 40 degrees, I have a much harder time getting into the ice bath. I have no idea why that is, but, but I liked that extreme and I also have experimented with like extended fasts and done like seven days with only water.

And that again, it's trying to push myself, see like how far I can go, how far out of my comfort zone can I go? And what does that, what do I feel like when I go through it? And so comes back to the meditation, comes back to like getting to know myself, getting to know what are my limits, what am I truly capable of?

What can I go through? And so that's why I love the ice baths and in Germany, it's super easy. Cause I just put a little kiddie pool outside on the it's so cold in the morning, it's frozen and I jump in.

[00:59:22] Ken: Oh, oh my goodness. Yeah. I still can't get to the super, super cold water. I it hurts too bad for me, but I'm working my way up.

And you know, that is a really, I think important point. And so every year from January 1st, my birthday, which is on the fourth, I do a water fast. And If anyone else is considering it, I talk to your doctor, do all the correct things. It's something that is I think it's important to have the disclaimer, but for me, that's the hardest thing I do all year.

Like food is my favorite thing. I love to eat. I love all food. I, you know, the highlight of my day is usually my meals, but giving that up is unbelievably difficult. And you also realize when you do that, like, if you can do that, if I can give up my favorite thing for the most part for three, four days, I could probably do anything.

Right. Everything else seems so easy compared to like jumping in the ice bath in the morning or not eating or doing some of these other things. And you know, if you really think about it, like, Okay, oh, it's hard for me to sit down and learn and focus. It's hard for me to do all these things like sitting down and learning data science is going to be so much easier than sitting there meditating while you're trying to think of just your breathing, right?

Like focusing on your breathing for an hour is boring as hell, right? I mean, obviously there's benefits and there is like, and there's other areas you go, but like doing data science is really exciting compared to just sitting there and only focusing on those, this like biological rhythm that you have. And so I look at all these, these areas of like difficulty or boredom as the opposite side of the coin that makes you able to do things that people might consider very difficult or that you used to consider difficult and you reframe them as being very feasible.

So that's an incredible way, I think, to kind of note, to end on are there any things that you're working on? How can people get in contact with you? How can people join the ML Ops community?

[01:01:28] Demetrios: Yeah, we've got, I mean, we've got weekly meet ups and podcasts, and obviously the Slack. If anyone wants to jump in there, there's all kinds of knowledge being shared.

And I'm super stoked on that. We besides the community docs page is also the like tooling compare page. That's taken. Probably a few years off of my life, trying to figure out what all of the different machine learning vendors do. And I interact with them every day. Right. But the the thing is. It's confusing out there.

And like, I feel for you, data scientists and machine learning engineers, we're trying to architect your stacks right now. And you go to a website and every single website looks the same book, as they say, like, we're going to help you productionize your model or 80% of all machine learning models fail, or like the data scientist spends 80% of his time cleaning data.

And so all of those things are super cliche and it's really hard to navigate this space. So I created, if you go to just MLOps.community, And you click on the learn page. You'll you'll see, we've created three thus far. One is all around feature stores. What feature stores do? Why are they different? What feature store is different than what feature store?

Another one is around monitoring. Cause there's like 18 different monitoring tools right now. And I have no idea what is different about them. And then I went and I did this study and I asked people to tell me, like, what's your differentiating factor? What's your value prop, right? How are you different than these other 18?

And we also did metadata management or metadata store, which is basically like experiment tracking and metadata store, metadata management. So we've done those deploy is coming up soon. Probably data labeling is coming up soon, but like I said, I'm really slow to go at these because it's just such a painful process to get it done.

[01:03:27] Ken: But credible value add still, I think. You know that that's what I would want. It's like, Oh, I'm thinking about using these tools. What the heck do they mean? Right. Like I, yeah.

[01:03:40] Demetrios: Yeah, if anybody has any like any advice on how I can make that more valuable, like, is there something right now, we're going to put up reviews so you can review different tools if you want and leave your review.

So you can hear feedback from others in the community, but if anyone else has ideas on what would be more interesting for you to learn about as you're going through that? I would love to hear it because that's super helpful to me. I'm kind of flying blind here. This,

[01:04:07] Ken: I love it. Awesome. Dеmеtrios, thank you so much for coming on the show. I'm really excited for when we talk again, and I think people are gonna really enjoy this episode.

[01:04:18] Demetrios: I appreciate you inviting me on here, man. It's awesome. Always a pleasure to talk to you.

3 views0 comments