Druid Summit 2024 Recap with #DataFemme and Larissa Klitzke

Transcript

Danielle: [00:00:00] Hello, everyone, and welcome to #DataFemme, where we engage you with stories of how innovators across the globe are using data to achieve new heights in their respective industries. I’m Danielle, founder of DiKayo Data. And right now I’m here with Larissa Klitzke, who is the senior product marketing manager at Imply.

Danielle: Imply just had an amazing event they put on called Druid Summit. So we’re going to be talking a lot about that. What happened at the summit and key takeaways and putting on live events and a lot of other topics that are key to data science and the open source community. So as always, sit back and relax and enjoy our episode.[00:01:00].

Danielle: Well, thank you, Larissa, so much for being here. I know that we’ve talked a lot back and forth and there’s been so much excitement pre and post Druid Summit, so I kind of want to start by just giving my audience a taste of what you do and what Implied does, so if you want to give a little bit of background there.

Larissa: Yeah, so my name is Larissa. I am a senior product marketing manager at Imply. My focus specifically is on Imply Polaris, which is our fully managed database as a service. And just for a little more context. Imply is a company that provides a commercial [00:02:00] offering powered by Apache Druid. Druid is an open source, real time analytics database that’s designed for fast, interactive queries on large datasets.

Larissa: It’s a lot by nature, but really it was designed for high concurrency, high throughput queries for both fast roll ups and aggregations. 

Danielle: Very, very cool. I have done a little bit of research into the whole Apache environment when I paired up with the developer advocate at Dremio. And he was talking a lot about Apache projects and open source.

Danielle: I know that Imply itself is a company now, but it was built out of an Apache project, which I think is so cool with open source being all the rage. Can you tell me a little more about that? 

Larissa: Yeah, so Druid was actually created around 2011 to 2012 at an ad tech company at the time. And the thing they were trying to solve was kind of the challenge around real time bidding, right?

Larissa: Where you have kind of these DSPs [00:03:00] and different demand side and supply side servers that are running thousands of queries on vast amounts of data at different sites at all times and trying to find that best bid on that best offer. And traditional data and data warehouses weren’t really designed to be able to facilitate that in the best way.

Larissa: So this was kind of their custom built solution. The original team that built it included FJ, who was our CEO, Vadim and Gian, who later founded Imply and Eric, who is our chief officer of emerging solutions today. And back then really the core mission. Was exploration, you know, subsecond execution of queries per streaming data and really having total flexibility of the data and types of queries that you could run on.

Larissa: So obviously that quickly expanded beyond just, you know, the initial use case of advertising and ad tech to, you know, really any industry. And it led to them actually founding Imply in 2015. And the goal with Imply was really to make a company [00:04:00] that could make Druid as easy for people to use as possible so that they could build really awesome data applications on top of that.

Larissa: So this includes enterprise-grade features, like support, added tooling and some extra bells and whistles that would not be included directly within Druid OSS. So that was kind of like where we started. And since then, obviously it’s been about 10 years now, a lot has happened. In 2019 Druid became a top level project under the Apache software foundation.

Larissa: And even back then the adopters were really big names, you know, think Cisco, Netflix, Hulu, PayPal, Shopee, Walmart, Yahoo, a lot more. And all of them are kind of building solutions that are powered by Apache Druid. And then I guess the last highlight that’s obviously relevant for me based on my focus is Imply Polaris.

Larissa: And that is our premium offering that launched in 2022, so it’s still relatively young, but seeing a lot of traction and growth. 

Danielle: That’s great. Yeah. Are you seeing the [00:05:00] same kind of traction as you did in 2019 with those same types of large companies? 

Larissa: You know, it’s a little bit different. I will say some of the biggest companies have entire engineering teams that are completely dedicated to Druid.

Larissa: So think like, Netflix, for example, probably would never actually use Imply directly because they are actually able to dedicate the training for that. But Druid, of course, is its own language. And it requires dedicated resources. So I’d say, you know, Imply Polaris is designed for companies of all sizes.

Larissa: But you know, if you become like a mega enterprise, we do sometimes see that they kind of spin off and go on their own. 

Danielle: It’s crazy. You know, when you say Druid is its own language, I feel like in tech in general, but especially data science, there’s just so much to learn. And you know, the next new thing could be so important for you to learn, but if you don’t have the time to just devote, you know, to learning at the same time as carrying out all your other [00:06:00] operations, it can be really helpful to have support in that way.

Larissa: Yeah. I mean, it’s, it’s not only easier, but it’s, I mean, it’s kind of counterintuitive for some, because obviously, you know, you see a price tag on a managed service, but we oftentimes see people that migrate to Polaris, either from Druid or other versions of Imply. When they actually do that cost time analysis, it ends up being cheaper just because, you know, you’re getting that managed service, and there’s a bunch of other add-on features like auto scaling and automatic upgrades that actually save a lot of time. And then obviously you’re, you know, you’re not spending that money on, on a team to do that themselves manually. 

Danielle: Definitely. And we will touch more on Polaris when we actually get to recapping your talk at Druid Summit, which is what we initially aligned to discuss.

Danielle: Very exciting. I was really drawn in as some of my audience might know already, but I was just so drawn in by. The lineup that you had and all of the talks [00:07:00] and when I reached out to you on LinkedIn, it was really cool to make that connection and coming out of the summit, what was it like? What were your main takeaways?

Larissa: But first of all, I will say, like, we had overwhelming attendance, a lot more than we were expecting, like, very, very packed rooms which, of course, is hard to predict, because, you know, you can do these webinars and kind of meetups and smaller events, and it can be hit or miss, but I think there really was a lot of enthusiasm from the community.

Larissa: This is also the first time that we actually did Druid Summit in person. Previously, we’ve always done it virtually, so it’s very exciting from that perspective, but yeah, I think there was just a lot of excitement in the room, and, you know. It seems like people were very well prepared, you know, they weren’t reading off their note cards and there was a good mix of technical and valuable content.

Larissa: So, yeah, overall, people were very happy and we did run a survey that confirmed that they were super happy. 

Danielle: What inspired you to take Druid Summit into a live event, you [00:08:00] know moving on from solely virtual to in person. What was the inspiration for that? Was it Polaris’s release or was it just serving the climate or I’m curious?

Larissa: Well, I’ll be honest. I’ve been at the company for just seven months now, so I don’t have the full context of the people that were originally planning it. I do know you COVID as to whether these events are worth it. And, you know, in some senses, like, we’re not, we’re not a giant company who has these, like, mega big event budgets.

Larissa: This is all about learning. But I think this also represented an opportunity for growth for us. But I will say it was mostly Druid users who were attending. So it wasn’t like a sales oriented event by any means. It was really about [00:09:00] fostering that sense of community. 

Danielle: It’s interesting how events that are more community focused are so different from the events where you can tell that sales is at the top of mind, you know, and I don’t, I don’t necessarily, I think both have a place.

Danielle: I enjoy going to both when I have the chance, you know, like I, I think one year I went to ODSC, the open data science conference on the West coast, but I had just come from DBT’s big coalesce conference that was in New Orleans where I lived. So I just got to go You know, commute from my house and everything, and it was so different, just the environment.

Danielle: Some of the companies do overlap, but you   just get the sense that the goals are different. And it’s, it’s, it’s interesting to observe how different events with these [00:10:00] focuses engage somewhat of the same user base in different ways. 

Larissa: Yeah, I mean, I think in some ways it influences the content as well, right?

Larissa: Like obviously we had, you know, a few keynotes and we wanted to talk about our roadmap, but really like 90% of the content at this event was not something we predicted in advance. We asked folks to submit sessions that they wanted to talk about and then kind of reverse engineered the agenda around that.

Larissa: So it really kind of came out as this kind of organic spread of a bunch of different cool things and learnings, tips and tricks. And kind of hacks of what people are actually doing with Druid today to build some pretty cool stuff. 

Danielle: Well, speaking of the content and, you know, kind of user generated content in a way, what was most surprising to you about the content that people came to you with?

Danielle: Were there trends that you didn’t expect to see that really came out of the woodwork? 

Larissa: Yeah, I mean, I guess I can kind of answer this from two sides. There’s probably like what I was surprised to see and maybe like what our attendees would be surprised to see. So I, I guess I’ll start with the latter, [00:11:00] which is really, I think probably what they were surprised to see were some of the kind of innovations in the roadmap, as well as kind of a divergence of focus, as I mentioned earlier, where we’re not just talking about Druid, but how Imply complements Druid.

Larissa: So some of those themes in terms of innovation include, advances in natural language query, some more movement towards supporting more complex SQL and BI workloads. We introduced our Dart query engine which is able to handle high complexity SQL queries at far lower latency than we’ve been able to do in the past.

Larissa: So it’s really just super powering what Druid is already known to form in terms of high speed, high volume. And then obviously you know, with the introduction of Polaris and actually some customers speaking about how they’ve used Imply there was kind of some discussion of, you know, how do people actually choose these solutions and kind of what are the trade offs or pros and cons of going with one route over the other?

Larissa: And like, why did they make those decisions? So I think it really kind of catered to like a range of maybe [00:12:00] people that are deeper into it and those that are still kind of in the exploratory phase I guess from my perspective, I wouldn’t say like there was one thing in particular that surprised me but I was just, I kind of mentioned this earlier, really just impressed by how prepared people were and you know, how much they had to say about how much they’d built, you know, we heard from some really, really big names like Roblox, Atlassian Netflix, Salesforce And they really went into a ton of depth about how they’ve actually built these things.

Larissa: And that’s something that, you know, you don’t get to necessarily encounter every day, especially since I’m not in a customer facing role. So yeah, that was interesting to me, but I wouldn’t say there was like, one aha moment in particular. 

Danielle: It’s interesting that Salesforce came up because I’ve been learning Salesforce pretty intensively and kind of unexpectedly in the past few weeks.

Danielle: And I just, you know, it, it makes me think about why data consumers are so interested in knowing [00:13:00] about the structure of how things work. And, you know, you were talking about there being a lot of content on how Imply fits into the mix and, you know, talking a little bit more about your role. I think that, you know, consumers of content, consumers of data, data practitioners are really interested in that because the documentation of how you process data is just as important as, you know, content about the results that that data might produce, you know?

Danielle: So the process and everything that’s going into developing how we deal with such large amounts of data. I mean, that’s relevant. That’s content in itself. 

Larissa: Yeah, no, no, for sure. I mean, it’s not something that most of us think about. But you know, obviously, like the vast majority of our attendees were developers.

Larissa: So this, this is what they think about, right? Because they’re the techies and the folks that are actually kind of spinning all those knobs behind the scenes.

Danielle: So you’ve been hearing [00:14:00] us talk about all things Druid Summit, and I’m sure that you want to listen for yourself to some of the sessions that you find the most interesting. The good news is you can. I’m here to tell you that you’re able to watch all the sessions from Druid Summit on demand if you visit imply.io/events/druid-summit-2024/.

Danielle: This link will also be available in the show notes so you can find it there, and I will mention it again at the end of our episode. You’ll find dedicated talks, not just about Apache Druid and Imply, but also about real time analytics, as a whole, what’s on people’s minds and what they’re trying to solve today from Netflix to Salesforce, Atlassian, Roblox, and more.

Danielle: You’ll find a ton of learnings about how you can optimize operations across databases, [00:15:00] analytics, data lakes, UX, and more. Now back to our show.

Larissa: I would say like most of the talks talked about how to actually optimize the data. And I guess, I don’t know, you didn’t ask this directly, but like, I’ll, I’ll talk about kind of the four key themes that we saw and how UX kind of diverged from that. So there’s four main themes. The first one was to do with entire data pipelines.

Larissa: And that was divided into lake house and streaming tracks which really emphasized Druid’s role in kind of bridging the gap between real time streaming and batch within a lake house environments. The next one was really about operations and optimization. So that’s kind of those data hacks, right?

Larissa: How to manage data retention and ingestion, optimize queries for speed and ensure reliability under load. So Atlassian, for example, shared some insights into their load testing strategy and others shared. You know, different methods of tuning the third one. [00:16:00] It was about advanced operational analytics with streaming and lake house.

Larissa: And this was really just kind of highlighting the relevance of data lakes and Iceberg specifically and kind of how that’s going to fit into the next evolution of data. And then the last one was about UX design. And this is really focused on how people are actually using Druid as well as Imply to build user facing and, or customer facing like visual analytics.

Larissa: So internally they might be doing this. If you’re trying to, you know, explore different metrics within your organization, you can think of this like, as you might use like a Tableau or a Grafana, for example. And then externally, those are software companies that are actually providing a service.

Larissa: And with Imply or Druid, you can do this in one of two ways. You can either funnel that data directly into a visual application that you’ve either built or a third party one. And the power of that is really making it way, way, way faster. I’m sure you’ve, you’re, you know, you’re used to, you’ve worked in B2B and you’ve used like a tool, like, like Looker, for [00:17:00] example, and you open the dashboard and you change a filter and then you wait like two minutes and you walk away and make a sandwich and come back.

Larissa: Drew has designed so that you don’t have to do that, right? Everything should be sub-second. But we also have a solution called Pivot which is included specifically from Imply and it’s. Basically like a build-your-own visual analytics application. Like you can just drag and drop and make it really, really easily without actually having to be like a developer and kind of do all of that UI work behind the scenes.

Danielle: That no wait time is really more important than a lot of people realize because. When you’re working, you know, you have this flow, you get in the zone and it’s, it’s hard to get there. It’s hard to get in that zone. So once you’re there, you kind of want things to keep moving. You want to test and verify. You want to see your results. 

Larissa: Yeah, totally. It’s funny, right? Because we’ve gotten so needy about just expecting a page to load immediately. Right. Whether that’s like on our phone or a laptop. And if it takes like 10 seconds, we’re like, Oh, what’s happening? But that is the world that I think like most BI [00:18:00] teams live in.

Larissa: And obviously that’s not to say that there’s not like a ton of value and kind of these bigger data warehouses, you know, whether that’s via Google or Snowflake or something else. I think if you’re, for example, in my previous role, I was doing you know, media planning in the first half of my career. And so, you and for that you might be pulling like a daily or a weekly report.

Larissa: And it’s fine, right, if it takes like a little bit for that report to load. But I think the reason that people really turn to Druid and Imply is for times and scenarios where you cannot afford to wait. So think about Netflix being able to provide an automatic recommendation on the top shows that are relevant to you based on some sort of AI model, right?

Larissa: Like, that needs to show up right away. Or if you’re, you know, a manufacturer and a plant goes down, you can’t afford to wait like two hours later. Or if it’s a bigger outage, for example, that’s affecting a lot of customers for their power needs. You know, that’s another real time scenario and you can think of a lot of those, I’m sure across like finance or healthcare or, you know, even, you know, applications, right? Which there are a lot of gaming and [00:19:00] more and so forth. So that’s really like where people are willing to kind of do that extra work to make it as fast as possible. And they might not want to settle for less. 

Danielle: And I think, you know, we’re getting to that point with our search engines too, where our search engines are so AI integrated and, you know, just optimized for that new technology.

Danielle: So I do want to hear a little bit more about Dart and NLQ and, you know, what that release is because that to me, that’s, just as exciting as, you know, the other inventions that you’ve been talking about, just because search is changing so rapidly. 

Larissa: So what is Dart? So Dart is a new low latency, high parallelism query engine.

Larissa: It complements Druid’s native engine and targets more complex SQL and data warehousing workloads. So initial benchmarks actually show that Dart can improve complex query performance by 2000% compared to the native engine for heavyweight [00:20:00] queries. And the way it does this is complex and something that I’m not, I’m not an engineer, so I can’t explain, you know, really specifically, but a lot of what Druid is doing to actually, you know, make things faster is it’s automatically partitioning data into different groups so that it can query from that fresh data set more quickly.

Larissa: So this is all done behind the scenes and it’s not something where you have to, you know, manually kind of move things around. So that’s Dart. It’s currently, I think, in kind of an exploratory mode and we’ll kind of see some improvements to that soon. And then Multistage Query, also known as MSQ, is not new.

Larissa: So it was released about two years ago. And essentially it enhances Druid’s capabilities by enabling complex data processing tasks through a distributed multistage execution framework. So same goal as what I’ve been talking earlier and same goals with Dart to make things faster. It just allows you to take multiple steps to kind of partition the process that ends up resulting in that kind of fast [00:21:00] query, but we did have you know, a number of folks that talked about how they actually leveraged that to achieve those results. You know, in different scenarios, which obviously is helpful to put that into context across a range of different types of companies and folks. 

Danielle: Yeah. Did you kind of notice a meeting of minds between people at different companies? Just passing by. 

Larissa: Yes. Yes, I did. And actually, like, I was somewhat surprised at how approachable everyone was. Like, there were a lot of folks walking around from like Apple, for example, right? And like people from Netflix and Salesforce that you know, these are these giant companies that obviously folks look up to but everyone was super approachable.

Larissa: And I saw a ton of chatter kind of in between the sessions as well as afterwards in the networking portion. So yeah, a ton of good discussion. I think, you know, people appreciate kind of seeing these different tactics and it’s, you know, not necessarily like in a direct competition way. That you might see with other types of topics.

Larissa: You know, my background is MarTech and my previous company was focused on mobile measurement. And so when we’d [00:22:00] have these kind of customer conferences, it was always kind of this interesting mix where people would want to talk about how they’ve done stuff, but not too much in detail.

Larissa: Because they don’t want, like, somebody who has the same type of app or service to copy them. But, yeah, I think it was great to see just how transparent everyone was about sharing that and how open it was. 

Danielle: I think that’s a really cool culture to foster, you know, because I’ve also noticed in marketing, you want to share these impressive stats that you have, you know, you get chills and shivers looking at certain stats that just paint you in such a good light, but then there’s always that worry. Oh, well, you know, if we, if, if our competitors know that doing things a certain way increases profitability by, you know, 200%, they’re going to do it too. And so there’s that balance right there.

Danielle: And I think It’s really cool that, you know, that you, something that you inspired in that culture, like, made people want to [00:23:00] share and treat it more like a, you know, I guess, bonding exercise instead of presentations that you know have an agenda. 

Larissa: No, for sure. And I mean, not to say that everyone gave like all of their secret sauce. I’m sure that everyone who presented went through a review internally. But I think when you’re talking. You know, tactically about how do you improve the performance of a particular application or a service? You know, that’s something that’s just, you know, a fun kind of technical hack to make it better.

Larissa: It doesn’t really go into the world of competition in terms of, you know, product positioning or actually like what that product is, or, you know, how you’re actually marketing to audiences for lower ROI and things like that. So, it’s competitive, but also not in a way, if that makes sense. 

Danielle: Yeah, let’s get into your talk, which was all about Polaris. What was it like to get up in front of an audience and talk about a technology that is quite new? 

Larissa: Well, first of all, it was fun to be on stage again. You know, in product marketing, I do a lot of kind of [00:24:00] webinar presentations and things like that. But I think the last time I went on stage was 2019.

Larissa: Just because of, you know, COVID and the, you know, obviously there being like less live events and such. But yeah, it was fun to get up on stage. I love speaking. I will say I didn’t realize until after I presented that all the founders were in the room, which is funny. So, you know, no pressure, obviously.

Larissa: But yeah, I think overall it went well. It was interesting because I think not everyone even knows what Polaris is. You know in terms of like what the actual product name is and like, how is it different? And I think the goal of my presentation was really to provide an overview of implied Polaris, but I didn’t want it to feel like a sales pitch.

Larissa: I actually wanted to go and kind of talk tactically about like, How was it actually different from other products? And I think, you know, as an end user, if you’ve ever been like trying to purchase technology, that can be really hard to parse out because you see kind of these sales decks and these web pages that talk about a ton of benefits and then you’re like, okay, but like, what is that actually?

Larissa: Right. And like, what is the Delta between what I’m using today? So some things I [00:25:00] talked about, I did kind of provide kind of like an overview of growth, and this was funny kind of as I was building it, because I wanted to, you know, normally if you talk about growth, you want to provide like a multi year growth chart, and I was like, oh, actually we have like less than two years of data actually in our system to pull this for but I did find that we had 12 and a half petabytes of data ingested last year and over 1.4 million logins. So it’s grown a lot in a relatively short time period. From a benefits perspective, I know I said not to talk about basic benefits, but really it comes down to ease of setup, ease of use, both from a backend and a visual UI perspective, kind of all the benefits of full management, cost efficiencies, and then kind of that security and resiliency as that kind of underlying layer.

Larissa: And then we actually went into detail and talked about a side-by-side comparison of Druid, like what, what are some things that Imply and Polaris has that Druid just does not focus on? And then what are some more technical use cases that you might not think of? So just like a few key categories.

Larissa: [00:26:00] Imply, obviously, like the managed services. That is not included in Druid. There’s some extra things around security and compliance. And then Pivot, our visualization tool, is something that’s unique. Outside of that, right, it’s like 50 different tiny little things. And everyone loves Druid, you know.

Larissa: You can do a lot with Druid in most of the categories in terms of, if you were to look at like a comparison table, there wouldn’t be like a binary yes or no in terms of Imply having something that Druid does not. So what I wanted to do was kind of pick out five unique things that are like technical use cases that were built and designed to be exclusive to Imply.

Larissa: So just a few of the callouts that I haven’t mentioned with things like elastic autoscaling which basically means that Imply automatically adjusts data as it’s ingested in real time. So if you think about the challenge of streaming data, right? If you’re, no matter what kind of use case you’re using it for, typically the amount of data use fluctuates on a somewhat predictable and [00:27:00] somewhat not predictable basis.

Larissa: And so you have this challenge of under or over provisioning. The amount of data that you’re actually paying to ingest because you want things to be fast, but you don’t want to overpay in terms of flow load. So that’s autoscaling, kind of like doing that all automatically. So things are faster and they’re cheaper.

Larissa: We also talked about dimension tables, which allow you to upsert data. So that’s like inserting new data or changing data automatically. And then I guess the last one that I think is maybe interesting is time series analysis. Which is, I guess, it’s own separate topic, right? There’s, there’s entire databases that are built specifically for time series analysis.

Larissa: Think like InfluxDB, for example. But this is something that we wanted to kind of build in to be compatible. With Imply and obviously, you know, come with all those other benefits, like real time visualization. Yeah, I guess like the main feedback, as I was talking to folks in the hall was just how clear it was.

Larissa: Like they actually learned a lot which was my goal, right? I didn’t want it to feel like a sales pitch, obviously, you know, it is just some extent. But they kind of appreciated, like, actually breaking those things down so they could actually kind of [00:28:00] have, you know, here are kind of a list of differences and they could kind of decide for themselves whether those, those things are worth investing in or whether they, you know, maybe want to try out Druid first and maybe see later.

Larissa: But yeah, it was great to see people seemed pretty happy with it. And of course I did like a little mini demo so people could actually kind of see it in action as well. 

Danielle: What was it like doing the demo? Is that something you’re familiar with doing? 

Larissa: I am because I recorded a demo a couple months ago. It’s online. There’s like an eight minute version. I will say it wasn’t at quite as smooth live as I was expecting it to because I forgot to switch from mirroring to, switch from extend screen to mirror. So I couldn’t see what I was clicking on. So I was like running back and forth to my laptop and the projector.

Larissa: But people are very forgiving. And if you watch the on demand version, we fixed that part. So you don’t have to painfully watch me running back and forth. But yeah, actually, like I was really comfortable with the demo. My main, like, concern going up there was whether I was going to have a [00:29:00] technical issue.

Larissa: And we realized the internet worked, but I didn’t think about that particular nuance. So. It’s just kind of funny with live events, you can’t plan everything. 

Danielle: Honestly, it sounds like you killed it. And I think it’s funny, I’ve seen a lot of times where the tech might take a little longer to get started at conferences.

Danielle: And sometimes that just helps ease the tension in the room for everyone, like take the pressure off. Do you feel like you got kind of a big picture of how all the talks fit together, too? 

Larissa: Yeah, I mean, overall, like, I will say, like, there was so much that was covered that I think, like, it was almost difficult, even just as we’ve been talking, to break it down into just a few themes.

Larissa: But yeah, there were definitely some categories, I think, that kind of pulled out overall, right? Things like, people that were more focused on that user facing experience and like how to actually build a visual application, there were, you know, how to actually choose a solution, I think was another theme.

Larissa: And then this is a broad category that includes a lot of [00:30:00] segments, but like, how do we actually, you know, make Druid faster and make it more efficient. And there were a bunch of different hacks that kind of went into that. So those were kind of like some of the themes then. Yeah, I think, like, from our participants’ perspective was interesting because we saw a lot of praise for the keynotes as well as the panels, which I find interesting, just because as somebody that’s attended a lot of events, sometimes I watch these keynotes and they’re like, very pretty, and I don’t get much substance from them, right?

Larissa: Like you’re talking about the history of, of data and the internet and people coming together and you’re like, okay, get to the point, right? But our keynotes actually were very info packed. And then I think the same thing with panels, sometimes I go to panels and I’m like, okay, I feel like you kind of just talked about a lot of things really high-level.

Larissa: And I don’t know if I have like, you know, a list of key takeaways to run with. Because sometimes it can feel fluffy, but I think just based on the folks that we had in the room and kind of the technical nature of it, we were actually able to get a lot of depth from it. 

Danielle: Yeah, that’s interesting too that there [00:31:00] you, you identified that it was a very technical audience because sometimes, you know, you get a mix, you get a lot of people who are more business focused and pushing the results, pushing, I guess, the opportunities that you find there at the conference.

Danielle: And then there are people who are there to learn and kind of up their knowledge of what’s the next thing that they need to learn. 

Larissa: Yeah, no, we had a few of those. Like, I, I mean, as somebody obviously that can see the attendees, like I would say it’s probably like rough math, like 90% or so developers or somebody within the dev org, and then maybe 10% of people that might be like a product person or a C-level person or some sort of executive, or a BI person that kind of works with those teams.

Danielle: Looking back on this successful event, like there’s so much you could say about why it was successful and what other events can do in the future. And, you know, for me, I’d love your take on why having these events [00:32:00] in data science specifically. As an industry is so important. 

Larissa: Yeah. I mean, I think there’s the obvious from like any events that you want to kind of foster community and discussion.

Larissa: I think that’s particularly true when you’re talking about an open source project like Druid, right. Where a lot of these conversations are happening online, but I’m sure, you know, you know, even from just working in a remote office environment that like actually talking about that live and having a dialogue is so important.

Larissa: It’s gonna be different. But I also think just because of the technical nature of it you can learn so much more from a day or even just, you know, a couple hours, for folks that actually wanna watch the on-demand content about these kind of hacks and these tricks. And that kind of inspires you to maybe build, you know, the next kind of big thing.

Larissa: And that’s something that you can’t get just from documentation. Right. So just kind of like a meeting of the minds and kind of learning from experts and being able to talk to them face to face, I think, is super invaluable. 

Danielle: I agree. I think that’s a great, that’s a great way to sum it up. Where do you think the future [00:33:00] of Polaris, the future of Imply is heading, like, when the, within the next year? 

Larissa: So overall, I would say in terms of like the future of Druid, like definitely check out the keynotes. There’s an intro and a closing keynote where they’ll actually, you know, you’ll see more about those specific features that I mentioned, but you know, it’s the same kind of mission and vision that we’ve been talking about all this time, like how do we make things faster and how do we make storage and data usage more efficient?

Larissa: Those are kind of the two pieces of the pie that I think Druid is great at. And obviously we’re going to continue to focus on that. You know, how do you do things really fast without breaking the bank? Because obviously, that’s the reason that people oftentimes don’t want to pay for real-time data is because if you’re querying all that data at once, it becomes very expensive.

Larissa: So I would say, like, definitely continuing on that vision. And then, you know, on the Polaris side, it’s, you know, how can we actually make that managed experience better? How can we provide further enhancements to you know, using that data, whether it’s from a technical back end perspective or that front-end [00:34:00] visualization.

Danielle: That’s what I was gonna ask. In terms like it is an open source project or like, there’s a lot of open source elements here, so are you looking for more contributions and like what type of people and backgrounds and types of contributions are you looking for or interested in? 

Larissa: Honestly the contribution happens organically. I forget the exact percentage, but I think it’s like 30 or 40%. Maybe it’s up to half of Druid is built by folks at Imply. But you know, there’s people like, you know, at Netflix, for example, that are regular contributors and they’re adding to the platform because it serves their own purposes.

Larissa: Right. So I think we don’t necessarily have to incentivize people to do that because they’re working and innovating on things kind of on their own. And then, you know, naturally that use case is going to be relevant for someone else as well. So in general, I think it’s been very, like, it’s very bottom up.

Larissa: And even if I think about kind of, you know, the way that we look at like a product product roadmap here you know, this is my first time working at a [00:35:00] company that’s built on an open source software and it’s very like organic, you know, we’re kind of building things as we go and then. You know, themes obviously emerged from there and we have some of our own vision.

Larissa: But it’s really, really defined by our customers. Like, that’s what influences everything at the end of the day. 

Danielle: Definitely true. What do you hope that the customers listening to this have taken away from the event, Druid Summit, and also listening to this  episode and this recap?

Larissa: My key takeaway is for anyone listening to this, whether or not you’re using Druid today, if you are building an application where real-time and fast analytics are important, definitely take a peek at some of the content. I would definitely recommend checking out the keynotes and from there, you can kind of explore and kind of pick the topics that seem the most interesting to you.

Larissa: If you’re using Druid, definitely make sure that you’re upgrading to the latest version. The latest version is 31. We actually had like a poll in the intro section of who was using older versions. And when we hear complaints [00:36:00] about things not working, the first question is, well, are you using the latest version?

Larissa: Or have you just delayed upgrading or you’re afraid to upgrade? So definitely make sure that you upgrade. Yeah, check out that content and, you know, create your own adventure. There are topics that are more kind of technical, things that talk about really specific topics like iceberg or lake house or, you know, natural language querying or if you want things that are like a little bit more high level you know, check out my Polaris talk, Tapcard actually did a really good job of talking about how they actually chose Imply and why. So there’s a lot to explore and yeah, hopefully folks get some use out of it. 

Danielle: That sounds great. I’m excited to go through and, you know, not live, but you get the idea, live tweet about the content. That’s usually how I keep track of things that I’ve learned at conferences is my, I know it’s, called X now, but my Twitter just kind of posting in a way that like inspires other people there to engage, you know, you [00:37:00] have your physical room and then you have the virtual environment that you’re also engaging with when you have a conference, I think that’s, that’s always been the most exciting part to me is that con, that content being a platform to continue networking and.

Danielle: You know, the networking being a platform to create more content through social media, right? Very, very give and take free flow relationship with those things. 

Larissa: Yeah. And I mean, it’s interesting, obviously like making the analogy again to the content itself. I mentioned this before, but we didn’t define the agenda in advance, right?

Larissa: Like the talks that you saw there, like 90% of them are just things that are coming organically. Like what are the problems and challenges that people are actually trying to solve today? From really big problems, like you see from like Netflix or Salesforce or Roblox to, companies that you may or may not be familiar with, like a Retailcloud or a TapCart, for example.

Larissa: So there’s really like a little bit for everyone and yeah, hopefully people can [00:38:00] learn something from it, whether or not they’re using Druid or interested in Druid, if you’re interested in developing analytics and making your data work faster, I think there’s kind of takeaways for everyone.

Danielle: Well, thank you, Larissa, so much. I really enjoyed talking to you about Jewhood Summit and all of the insights that you had there. I hope to continue the conversation on LinkedIn, on our socials. We are DiKayo Data or DataFemme. You can find everything pretty easily and we will be posting about this episode of course.

Danielle: And just again, to everybody listening, you can find all the sessions from Druid Summit on demand at www.imply.io/events/druid-summit-2024. We will see you next [00:39:00] episode.