Ask Anything, Solve Everything with Altair

Speaker A: Hello everybody, and welcome to the dedicated show. My name is Kate Sinachni. I'm the founder of Dedicated and your host for today's show. I'm really excited because today we are talking about Ask Anything, Solve Everything with Altair. A special guest that I'm going to have on the show in just a minute is Dr. Ingo Mjersva. He is the founder of RapidMinder and currently the SAP of Product Development at Altair. Before I bring Dr. Ingo onto our virtual stage here, I just want to remind everyone how this works. We are live, so if you've got questions, if you have comments, please go ahead and put that into the chat. I always encourage people to let me know where they're tuning in from. Introduce yourselves. This is a great way to network remotely. We all live across the world and I think it's 9am here for me Eastern time. And let us know what time it is for you. I know for Dr. Ingo it's about 3:00pm for him. Um, I love the fact that we all get to not only talk to each other here, uh, virtually, but also you guys can chat in the comments and build some connections. I'm going to go ahead and bring Dr. Ingo here on to our virtual stage. Hello. Hello. Welcome to the dedicated show.

Speaker B: Hey Kate, Good to see you. Good to see everyone. Um, yeah, 3:00pm as you just said, I have a couple of wine glasses here, so maybe it's a little bit too early to get the first glass of wine. Let's skip that for later.

Speaker A: But yeah, after the show.

Speaker B: After the show. Maybe that, maybe that.

Speaker A: Yeah, absolutely. So I think a good place for us to start is, you know, talk about your role at Altair. Tell us what is Altair for those who are not familiar, um, and talk about the rapid Miner acquisitions, basically. Tell us your story.

Speaker B: Sure. Well, let's start with Altair. Altair has been around for a while. Um, it's 35, almost 40 years, uh, on the market, starting on the simulation side. So really more on the engineering side of the world. Now I personally did study physics, uh, next to computer science in university. However, I would not claim to really be an expert on that side of Altair, but it is fascinating. So a lot of the use cases are really around like physical problems, like simulating, for example, the impact of crash on, let's say on a car, how it's going to deform, where the different forces are, uh, going to work. And sometimes I look at those problems and really think, feel like, oh, this is a Great area where machine learning and AI can also help. Which is also the reason why our founder and CEO, um, Jim Scaper really talks a lot about the convergence about three different areas. One is the world of simulation, one is the world of AI. Bringing those together and then both are backed by the third leg of the stool if you like, which is our high um, performance computing area. Um, those have been a couple of spin offs out of NASA for example. And uh, we really kind of like can claim that pretty much all the supercomputers this planet run on Altair's HPC solutions. So you have like super strong compute, uh, you have AI, you have the whole simulation engineering world. And the converse of all of this is really what describes Altair best. Sometimes we call it data science meets rocket science. And that's kind of.

Speaker A: I love that.

Speaker B: Yeah, yeah, so, and yeah, you asked about like how. Well the Rapid Miner story event, um, so as RapidMiner before we got acquired, uh, about two years ago at this point, um, we obviously have been focusing completely on data science in general, um, but mostly on machine learning models, building them, testing them, deploying them, uh, then actually putting workflows and decision frameworks around this. So this is kind of like what we today are really meaning when we think about like AI based solutions or even agents. Um, those are things we have been doing pretty much like for all the time. So we built this whole platform around this um, called DO Rapid Miner platform. And then uh, yeah, uh, we had a very, very strong user community, more than a million people on the planet using the, the now Altair Rapid Miner platform, um, and a lot of enterprises. So we've been very, very successful and yeah got acquired like two years ago and now it's really bringing those two best worlds together. So we have still the vision and the technology and the great team of Rebel Miner, but we merged it really with another great team which uh, has been with Altair obviously already before. But I also would just call it the sheer force of Altair. I mean it's a complete different league really. Like you have like, oh my gosh, you actually have like a complete marketing department and thousands of salespeople and all of a sudden like a lot of things which are always difficult or more difficult as a small startup become easier to be honest and just scale at a complete different level. So which is awesome for us, we have like more people or we merged with an existing engineering team on the Altair side as well. So kind of like double the resources. They alone. We uh, 10x the resource for customer success and customer support, which is another important area. And all those things are like, you can only dream of if you, if you're like in a startup world, really.

Speaker A: So, yeah, I was going to say, like, what was going through your mind because you were at Rapid minded for what, 10, 15 years? I feel like. Yeah, yeah, it's a, it's a long time, right? That's, that's someone's like, big career. And like how, what was your reaction? Were you bought in right away for this acquisition? What was going on?

Speaker B: Yeah, yeah, no, it's like, uh, uh, I, I, I, I. We've been talking to each other already for a couple of years. So most acquisitions, they are not kind of happening overnight. So you build a relationship, it takes some time and it's, it's, and Altair always have been, has been like, kind of like a, a dream partner for, for ours. Um, and so because I also sense like, similar vision, similar culture, um, the people here are amazing, um, really incredibly smart. And I think all of this matches very well. I think this is the most important thing. Um, the other element I liked about Altair, Altair, um, in general is pretty acquisitive. And so it's not kind of like the first time around. We had a couple of other candidates who have been interested in acquiring us. I obviously can't go into the details, but sometimes you feel like, yeah, m. This would be like the first major acquisition for organization. And are they really knowing what they're doing or are we ending up kind of like a bit in a, in an empty space? And, and, and that can happen. It does happen. So when Altair was very obvious, there's a, there's a clear strategy around acquisitions. Um, so what happened really like from day one, literally was like we ripped almost like everything apart and merged into the organization. So everything was completely integrated within like three months. That's true for like processes, teams and everything else. And that is kind of unheard of. I find this being acquired by other companies, it's like it can be years and like, wait, you're on the same teams instance already? Like, of course we are. That happened on day one. So all those little things, they really helped. So Altair knew perfectly what, what, um, what the organization is doing. But then the same is true for like the technical integration and we can go obviously deeper on that one as well. But, um, same is true there. So there was like a complete buy the, into the RapidMiner vision. So it was very clear that, um, RapidMiner became kind of like a central part of the overall platform. And so that is a great feeling for you as a founder as well. Obviously, if, you know, like, okay, it's not like your technology gets acquired, your team gets acquired, and then maybe something good happens, you start to realize there's a great bright future ahead. Um, so all of this was fantastic. So, like, yeah, I'm sure other acquisitions would have been working out as well, but this one was spectacular to say that in a completely unbiased way, actually.

Speaker A: So, like, yeah, nice. Well, I'm very glad that everything went smoothly. This is a, uh, great acquisition story because, like, we both agree that, uh, it doesn't always happen that way. Sometimes the company buys another company and they're just like, oh, well, let's just let it sit and die a slow death somewhere on the side. And sometimes it's a beautiful partnership. So I'm really glad to see that. And you look very happy as, you know, a founder who's now part of Altair. So I'm glad to see that.

Speaker B: Absolutely. I mean, like, yes, there's another element which, which is interesting, uh, since ate are quite so many companies, um, there are a lot of founders. Like, it's almost feeling like, like every, I don't know, 10th person you meet is another founder. Like, oh, you founded this and you found that. So. And, but it is an interesting, very innovative, very, uh, kind of like engineering driven culture, if, if that makes sense. Like it's very open for additional innovation just because there are so many people a little bit like that, um, running around here. So I think that's, that's really a huge advantage. Um, and that's another thing not every other organization can offer. So.

Speaker A: Yeah, yeah, definitely. So I'm just checking into the comments. I see we've got folks from Ohio, from Texas, from India. Um, Alex, I don't know where you're from, but you're saying hello. Hi, Scott. Hello. Hi, Robert. Thank you guys for joining. If you do have questions for Dr. Ingo, please make sure to drop them into the comments. This is a great opportunity to ask anything. We're going to try to solve everything without terror here. Um, so Ingo, you were talking about acquisitions, and I was reading that recently there was an acquisition of Cambridge Semantics. So tell us about how does this actually benefit your customers? What does Cambridge do? And yeah, just tell us that story now, because now you have. Do you have another founder in the company now?

Speaker B: Yeah, yeah, actually a whole bunch of additional founders, which is awesome. Like many other organizations. This one too, it's multiple uh, things founders. So yeah, Cambridge semantics is really all around knowledge graphs and um, giving data what I would say call like a semantic meaning. And um, why this is so important. Um, well, I think personally it always has been important, but there have been a couple of reasons why knowledge graphs have not been that widely used in the past. It can be sometimes, um, challenging to create what we would call an ontology. An ontology basically describes the world. Okay, um, so you have different objects in the world, they interact with each other. You have different people. Like there's stupid little examples like, hey, there's actors and an actor plays in a movie and a director directs a movie and a producer pays for the movie and stuff like that, whatever. So you have kind of like the different entities and how they relate with each other. If you have this knowledge about how your, the world works, this knowledge can be incredibly useful. Um, also for like data science, machine learning, AI. And now the funny thing is a couple of years ago generative AI became a much bigger thing. Then all of a sudden everybody started to realize, huh, huh, this is ironic. We kind of trained large language models on, well, the language assets we produce as humans. So they tend to understand us humans better than those models would understand, basically, let's say machine data or data in general, because they have not been trained to talk like that. It's a bit of like an ironic thing really if there's kind of like computer based model and it doesn't really understand computer based data. So, and this is, this is a bit of like a disconnect. Um, and we start to see this pretty early. This is one of the reasons why we got so interested in knowledge graphs that like, if you want to help, um, generative AI models, for example, language models, uh, we would need to basically provide a description what the data is actually about. If your table is called XYZ1 2 3, it's meaningless for you as human as well as for any generative AI model. If your columns are called abc, same story, you have no idea what's going on there. But if you have an ontology on top which describes the semantic meaning of everything, hey, this table describes our customers. This table here describes our transactions. This column here is actually the customer ID which points back. So yeah, you can think of this a little bit like in a relational database scheme, but knowledge graphs allow you to do that for pretty much any structure. And you don't. You're not kind of like confined to the world of relational databases. You can include unstructured data, you can basically have like much more complex setups. Like the world is just much more complex. So we believe Knowledge Graphs become an incredibly critical um, piece of every future AI stack. Um, and that's one huge reason. Well the other one obviously the team is just fantastic. The technology is rock solid. It's um, massively paralyzed, incredibly fast. Especially, especially for analytical queries. So all of this was very intriguing. And then like when we looked deeper and deeper, this also created then a complete new vision for our altered minor platform where we can now take the best of the concepts of a data fabric and the whole AI platform and bring them together. So we call this whole thing then the AI fabric because it's really what it is. It takes those different elements of those two worlds and also makes sure that the AI fabric then can become a part of every single business process. So and all of that would not be possible without knowledge Graphs. So yeah, that's why we m made this move.

Speaker A: Yeah, thank you for sharing that. So I want to understand better. So how are your customers now using Rapid Miner Altair and the Knowledge Graph tool? It's called Altair Graph Studio Now.

Speaker B: Right, right, right.

Speaker A: Embedding the Cambridge.

Speaker B: For whatever strange reason, we love studios. We have uh, what used to be RapidMiner Studio now is AI Studio Studio now. Our graph product is controlled by Graph Studio. So we have a lot of studios. Um, I think we also have an IoT studio and at least a dozen more. So yeah. So the product with which you basically design the ontologies and work with the graphs called Graph Studio now and then there's the Graph Lakehouse underneath and that is now deeply integrated with the rest of the AI cloud which is pretty much one of the cornerstones of the uh, ltm, uh web liner platform. Now um, so question was like how do our customers benefit? A great example in my opinion is um, uh, say a leading pharmaceutical company. Um, I can't name the name. I guess at least I am, I'm not sure if I am allowed to say the name so let's maybe skip the name. But a huge pharmaceutical ah, company and um, what they are using basically a combination of those two technologies for is to create kind of like a co pilot light experience but on their data and that data is incredibly complex. It describes clinical trials. You have all kinds of information about what has happened in the past. You would like to actually ask questions like what was tried and how, et cetera, et cetera. So normally this would be for every single question a business user would need to ask. It would be A huge effort to actually even figure out where's the data, how to bring it together, and to get action to any form of answer. Now, here's the beautiful thing about knowledge graphs. If you describe this whole world, the ontology about clinical trials, if you have all of this in place and the data is in the graph underneath this ontology, then you can actually, um. I'm not saying it's easy. Uh, so that's why we have the whole AI platform. But you can use the AI platform on top of both to actually create a conversational system which is not like a chatgpt, like experience where you just ask stuff about like, hey, give me your favorite apple pie recipe. But you can ask directly questions about the data, which can then translate it into what we call sparkle queries, which run against the database, which produce then the, um, answer. And that answer then is not hallucinated or anything because it's coming straight from your data. So bringing those different areas together really is something where customers say, like, whoa, whoa, whoa, whoa, before every single question. And I have, like 100 every day takes, like, I don't know, at least one hour, two hours for some data engineer to sit down, actually figure out what's going on, if you're lucky. And now I basically type it in. Boom, here's the answer. Okay, great. Now let's do the next one. Boom. Here's the answer. And that level of iteration and speed is something which, like, blew their mind. So we rolled it out for like, 10 people. That only lasted for, like, two weeks because the rest of the organization wanted it. And now it's going to be rolled out for thousands and tens of thousands of people because it's just, it's a complete game changer. We have never seen anything like that before.

Speaker A: That's always great when you do a small rollout and everyone's like, me next. Me too.

Speaker B: Exactly. Yes, we like that. Yeah.

Speaker A: So it's funny because as you were talking, I wrote the word hallucinations and then you addressed it. So you said there, there are no hallucinations because it's based on your own data. Right, exactly.

Speaker B: So you could theoretically produce, uh, a, uh, form of hallucination, which could still happen, theoretically. I mean, don't want to go super deep in the technology here, but, um, you could basically hallucinate in the sense of like, hey, I believe there is a different part in the ontology, although you never showed it to me, or I believe the sparkle query should look like that. But, um, it's not actually how the data stored in the database. But those hallucinations are easy to catch. So if you basically would do this like a mistake already by making hallucination about the ontology, you would get the result like a sparkle query back and you basically test it against the ontology and you would see immediately without executing anything, dude, that doesn't even run, that doesn't even compile, that doesn't even, that doesn't even match. So I'm not giving you any answer because it would be the wrong one anyway because that thing made something up. So it's very easy to catch those hallucinations. Same is true if you would produce some sparkle, which would not execute, which pretty much never happens. But if it does happen so again before, before you actually would make a mistake, we can catch those mistakes. Um, but as soon as something is valid against the ontology, that's the beauty. We know it's then also completely backed by the real world. So like that query is a valid one. It may still answer the wrong question. That's a different thing. There may still be misunderstanding. But then, then the query is executed, it runs against your data and you get like the true result, not just like 42, because the model thought 42 is a cool number.

Speaker A: Right?

Speaker B: So, yeah, that's, that's kind of the difference there. So like the output is not the, uh, generated answer, is not the immediate answer. It's basically the query which then gets executed to deliver the actual answer. That's okay.

Speaker A: Okay, so but let's say, let's say the answer was 42 and it was right. Um, but you wanted to validate it. Is there a process in place, like an audit trail that you can see? Like, where did that number actually come from?

Speaker B: Yes, uh, there's actually multiple layers. So the first one is basically an exponent explanation of what actually has been computed.

Speaker A: Okay.

Speaker B: So because not every person speaks Sparkle, and although it is easy to read, I would even claim sometimes even easier to read than SQL if like in other languages, query languages, it's still, it's not for everyone. So there is like a textual description, what actually has been executed, or if you want to go really down one level, then you can say, you know what, show me what actually has been running. And so there's kind of like almost like a debug mode if you start wondering, like something, I don't know, at 42, really, I don't think. Huh, huh. Let's see what was actually done. And normally that's what I said before, what can happen? It's Pretty rare. But what can happen is that there's just some misunderstanding in the, in the question. So like you would ask like uh, I don't know, how many trials have we done for, in, in, in this geographic region? And then for whatever, let's say whatever Emea, for whatever strange reason. Now the, the calculation, what actually encapsulates Emea may be done in a different way or like, like okay, maybe there's two different kinds of trials and it's pick one and should be the other. But in all those cases you would actually still get the explanation. You would immediately know, no, no, no, that's not what I meant. And here's the beautiful thing, it's still true conversation. You can basically ask back, no, that's not what I meant. I don't want Emir like this, I want Emir like that. Or I want the other type of trials. And then it would correct the query again and you would get the, with the different results.

Speaker A: So just like when speaking to an engineer who's pulling the data for you as well, right. They might have clarifying questions like what exactly do you want? Except it's much faster.

Speaker B: It is. And that's exactly what it is. So I would not claim to say like, hey, it's always going to be guaranteed the correct answer. Like I would never say that about any human being. I mean like it's ridiculous to ever expect that from any machine learning model really. Like that's not what happens really like but you get it with an instance and plus you get the explanation and then if you don't like it and this is a huge advantage over the state of the art, like two years ago, now we can actually have a conversation about fixing it versus oh, now I need to open my Python ide or now I need to open, I don't know, my alter, uh, reprimander workflows to actually make a change. Now all I need to do is like no, that's not it. I want the other trials and that's the way how we're fixing it. And that's a, uh, complete. Again it's not just flow. Speed is also like about democratization because all of a sudden everybody can ask any question. All you need to be able to do is type or speak. Actually if you have like some speech to text. But that's right, not really necessary.

Speaker A: You know, I think it's so helpful when um, when people share examples like you just shared that pharmaceutical example. Do you have another customer story without revealing names, um, that you can share, maybe a different use Case of how Rapid Miner and Altair technologies being used. Like one of your favorite stories.

Speaker B: Yeah, well, one of the favorites, it may not be the biggest one. I mean, like, there's obviously hundreds and hundreds, but, um, one I really like because it also would probably never have happened without this merger, is like, uh, a huge, um, international paper manufacturer. Um, and it's obviously a huge industry. I mean, despite all digitalization, we still pretty much use paper. Exactly. So like, it's not going away. Like, uh, there's things called books. I heard they're still existing. So.

Speaker A: Yeah.

Speaker B: Anyway, so obviously huge industry and paper gets manufactured. There's like machines involved in everything else. And it was actually one of our, uh, data scientists, uh, not a computer scientist and not like trained like that, but has been using Rapid Miner for I don't know how long. And when we became a part of Altair, um, Scott, that's his name, came came up with this vision of like saying, like, look, um, here we actually have the chance to create what now is called the AI Edge device, which pretty much like is a small device with the camera and network connector and everything else. It's a tiny device and the whole altayrab Miner platform has been packaged in a way, so it's actually running on that device. It's a small edge device. So since it has a camera, we can actually put it next to the production process, in this case for paper, and actually can take pictures from whatever is going on. And there are, for example, situations where, uh, some of those machines, they kind of clock up or there's some, some dirt. And you need to basically pay attention to this to see what's the level of this. And if it's too high, we should probably do some maintenance there. So, classical predictive maintenance type of application. Nothing too crazy. But the twist here is instead of taking some sensor data, whatever, like, you don't need to do anything. Like, as long as you have some space where you can plug the magnet with the device like somewhere there, or put it on some tripod or like get it like screwed next to a wall. It doesn't really matter as long as you can point the camera, uh, produce a couple of training examples and do it that way. The beautiful thing is it's running on site right there. So it captures the images. It basically applies the models. It does everything, triggers the decision and basically turns on a red light like next to this, uh, uh, uh, machine, which is basically checked and measured. So why I love this for so many reasons. A, it's an engineering customer um, the second element really is we have the area here talked about a lot of tinkerers and thinkers here, and we try to experiment and do things, but we also have the space and the freedom to actually do that. So this thing, which started as one person's dream, Scott's dream here, really turned into, uh, a complete product. Uh, and we just rolled it out to a bunch of other clients because you have tons of use case, put the same device and see if everybody's wearing their hard hats. And so it's important for the security of the employees as an example and so on and so like, but it's all packaged up so you don't need to really figure out like, oh, how I'm not doing this. And the most beautiful thing is there's not a single line of code you need to write for anything. So it's kind of like you open up, you plug the device somewhere, then you have like little Lego blocks you move around. I don't think I can call them Lego blocks, but whatever little blocks operations which are, uh, which have to be moved around to define what's actually happening and how everything works. But everybody can do this. You can customize it to your needs and then, uh, yeah, you have kind of like machine learning on the, on the shop floor. So it's pretty cool.

Speaker A: I think I just thought of like a hundred use cases in my head while you were describing that. I mean, from a safety, there's just so many use cases.

Speaker B: Exactly.

Speaker A: That's crazy. And I'm actually sure that we're going to see those pop up everywhere.

Speaker B: I, I hope so. So I'll tell AI Edge device. So if you see it like everywhere, like, like, don't get nervous. They're going doing good stuff. So like, even if the camera is pointed at you and now it's something like, like it is at me right now. I, I think this is, I mean, like, I don't want to get to like philosophical, uh, now, but this is just like what still excites me. I'm like, I'm in this field since 25 years. And it's like, yeah, sure, there's a lot of ethical questions and we should pay a lot of attention to those. No question about that. But it's those really, those new opportunities, those great use cases like the copilot, like experience on your own data, like without hallucination we discussed before. And all the AI Edge device, those are the things which really move the needle, which really change how organizations are working. And I still find this exciting. Every single day.

Speaker A: So, yeah, it is very exciting. Especially I'm an optimist. So I'm always like, oh, this is going to end up in the right hands. We're going to solve all the world's problems. Problems. And um, hopefully that is what ends up happening. But you're right, there are for sure ethical considerations that need to be thought of and thought through and that might hinder the speed at which we'll see all of this implemented. And that. I think that's necessary.

Speaker B: That's. It's. I agree with that. Um, and to be honest, since most of our clients are larger enterprise anyway, they want that from us too. I mean, like, it's. I'm not, I'm not dismissing this. It's just like it doesn't excite me as much. It excites me more to think about the optimistic, great new world than like, oh, yeah, sure, now actually let's slow down again a little bit. But it is necessary and it's the right thing to do, so no question.

Speaker A: Yes, of course. Um, okay, so Altair was recently named a leader in Gartner's Magic Quadrant for data science and machine learning platform. So huge congratulations here to you guys celebrating. I love it. Yep. What do you think caused, caused you guys to be placed in the leader quadrant?

Speaker B: Um, yeah, so I think first of all, we, ah, are incredibly happy, um, because it's a huge achievement. Um, I would be certainly also incredibly happy if we would not have been a leader but somewhere else on the whole MQ thing because, yeah, not everybody reads those. But for many organizations it is an important tool to actually inform themselves about the market. I always actually encourage to read and not just look at the picture. There's also another document which is called Critical Capabilities, which I really, really highly recomm. Actually much more than the MQ itself. But I don't want, yeah, but I don't want to dismiss obviously our success. So we are still, again, did I already say who, who if not. I'm saying it now again. Yeah. So now you have to question. So, yeah, like, obviously, um, in Gartner's terms, um, we showed like a way stronger vision as well as a way stronger execution than most, uh, of our competitors. And I would say that's true. So we always have been with Redminer before, very strong on the, on the vision side, um, since many, many years. Uh, we have been also leaders before already and it was never really the vision side. It's. We just have been in the incredibly lucky situation that we have Bright people who have been thinking forward and actually also managed to really get uh, some of our clients excited. And that excitement then like was transported over to Gartner. So they all always realized that. And so like this was not too much of concern to be honest. However, as a small company versus like some very strongly funded organizations, the hyperscalers and everybody else, uh, just in comparison, it's hard to keep up on the execution side, to be really honest. So that really truly changed with Altair now in the back. So uh, yeah, we use the whole Altair RapidMiner platform. As uh, I said earlier, we use the RapidMiner brand but it now includes also a lot of other fantastic products they contributed, uh, because it fixed a couple of um, areas or added a couple of functionality, not a couple of. Some functionality we didn't offer before. So that is not obviously kind of like a one plus one equals three situation that like, oh great, look at this really strong data realization layer here. Or we can now extract information from unstructured data. We have um, other differentiators like uh, for example our SOC products, which are also part of the Altair Rapid Miner platform. They allow us to um, offer like an alternative for code, uh, written in the SAS language. So you could basically run SAS language code, um, with this alternative compiler, uh, as well. So like all those are strong differentiators which now came all into the mix and that's fantastic. But the other elements keep always in mind. Gartner is not just looking at the technology. Um, actually one of the Gartner analysts at one point made a comment which was really going along the lines of like, look, if you would just look at the technology, RapidMiner always would have been a leader, always will be a leader because it's just an absolutely fantastic platform and I'm really, really glad and appreciate that. But it's broader than that. It's the whole organization. How well can the organization support you and be a partner of their clients?

Speaker A: Right.

Speaker B: And again with more, with more people and larger power, kind of like in the back, this is easier than it ever was for us. But another area of the organization here is um, what we call our business model, but it's called Altair Units. Um, so with Altair you're not buying a single piece of software and then use it and then, oh, you want to use a different software or buy it again, here's another subscription. That's not how we work. So the way how we work is you buy what is called Altair units and then while you're Using any of our products, those units gets checked out and then when you stopped using it, they go back into your units pool. So if you ending up using multiple products in parallel, whatever, you can do that if your units pool is large enough, kind of like a maximum bandwidth of products you can use, but you're not kind of like settling on a specific product and then forever. And sometimes, especially in data science AI, we have situations where like, hey, let's do some data prep. You do that for a couple of weeks and then you end up like, I don't know now you do some modeling, you do that for a couple of weeks and then you end up like doing a lot of visualization work and you do that for a while and then actually end up doing nothing of that because you're tingling around like, like you're presenting the results like to other people. Four weeks, no product use whatsoever. Normally with a named user subscription, you would keep paying for all the stuff all the time. So the total cost goes up and it's really not very flexible. Again, with Altair units, you just stop using it and the same units could be used by your colleague. So it's really kind of like, it's a very flexible, very powerful and very affordable way to buy software. It's something I've never seen anywhere else. It also makes it very simple, um, for both us as well as the client. It's also another reason why the acquisition was so smooth, because from one moment to the next, like as soon as we also implemented the units in system into the Rapid Miner products, all of Altair's clients got access to the Redminer products as well. So we did not need to go through additional paperwork and like sales processes and all of that stuff. So it's, it really accelerates all of that. So it's a patented model. It's a very special model. But again, as I said, it's, it's everything. It's the product, it's the vision, it's uh, who we are as an organization. It's the center of excellence, the educational areas, or it's even things like the business models. So all of, like the units, all of those go into something like an mq and um, then explain why we ended up as a leader there.

Speaker A: Yeah, yeah, I absolutely love. I didn't know about this model. So I'm, um, glad. I just feel like it would enhance so many benefits like collaboration where unit business units would work together. Hey, we use this now. You can use this. We're not using this anymore. Scalability where you use what you need when you need it. And it just makes so much sense. And you can also, you don't have to buy a new tool just to try it out. You could try it and if you need it you get I guess more units, uh, for your pool or whatever. However that works. It's a brilliant model. I love it.

Speaker B: It is a pretty model. I like when I learned about this the first time there was actually pre acquisition I was already like that is really, really, really smart. And then I wonder like why is not everybody doing this? And then I realized like there is actually a patent on that model. So that's why uh, Altair really is very unique in that regard. I mean it's really the best of like a consumption based, value based model. Like you have it from cloud providers, uh, you have this but you, so you have the flexibility but at the same time you have like a lot of control over the cost because like you know what your units pool is. So like you can't over consume really easily. So it's, it's also like if I would be the CFO of our clients, I would love that because all those benefits we just discussed. But I also have, I can sleep well because it's not like I'm waking up next month and like what is our AWS bill now like? Because I. Ah. So yeah, so it really combines the best of all those rules.

Speaker A: Yeah, right. It makes, it makes everyone in the company more intentional with what they're going to use and why they're going to use it and okay, I love that. M. So a couple more questions for you. So one is for, for people who are interested in this, like how would you recommend they actually get started? Do they get a unit? Is there a trial? How does this work for, for Altair? For Rapid Minor? Yeah.

Speaker B: Yeah. So like uh, well we kind of like have our own uh, uh, well if you like clouds or area, uh, Altair one. So like you could basically log in and there's a lot of products where you can basically just start. There's free versions of some of the products as well. So um, all of that is available as well. Um, one of the best ways in general to get started is sometimes not even with a product, but is with some form of education. A lot of our products are um, complex. I mean AI, machine learning, data science, those concepts are still sometimes hard to understand. If you're new to the field, you should actually learn about the concepts first and not just what buttons to press. And the same is true for a Lot of our other more engineering focused products, uh, uh, like from the hyperworks platform or from our HPC platform. Um, they are kind of like technical products and you should learn about them. So at minor ready, same now with Altair, we invest a lot into education as well. Um, so one great way of getting started is visiting our academy. So there's multiple places where you can do that. There's learn.altair.com there's also the academy.rapidminer.com I think the URL still works, I, I hope. But um, uh, if not just Google for academy and revenue, you will find it. Um, and there's like, I don't know how many hundreds of hours of like video materials which are basically explaining you the concepts of machine learning, data science, AI, how to solve complete use cases, um, and really on different levels. And I think this is the best way to start in many cases, especially if you're new to the field, if you're already an expert, sure, just grab the product and try it out. But if you're not, um, or if you would like to get a bit more guided approach, the academy is a fantastic way of doing that. So I highly recommend this um, as a great starting point or at least as one of the good starting points. And then yeah, there's obviously the website and social media and everything else as well.

Speaker A: Okay, perfect. And I'll make sure to put the links into the comments after we wrap up so people can click on them and make their life easier. Um, we did have a question in here earlier from Rabid. He was asking what's one thing you love about Altair? Just pick one. And one thing that you find different from uh, your startup world.

Speaker B: Yeah. So I think the second part of the question I kind of like, uh, answered already, uh, before and I would just quickly repeat it. Um, so in the startup world it's always, you're always kind of like fighting for resource. You're fighting kind of like for everything really. It's always there's some struggle, it's a lot of fun, I love it. But you're always constantly struggling because there's more customers, more. If you are half successful, I think it's a different kind of struggle. If you're no success whatsoever, then you're struggling with the fact that nobody actually wants anything from you. That's not great either. Exactly. You're sitting there, great, I raised like 10 million in venture capital and now actually I don't know what to do with this because nobody wants anything from us. So that would Be a different kind of struggle. Luckily we have never been in that situation really. So, so good. But like at the same time it's like, well, if you have like a million users and like a lot of clients with like 250 enterprise clients, like, there's always like, sure, can you also do this? And this needs to happen and that like everybody wears every single hat possible. And so like as much fun as it is, it is also demanding. Not just like for you as a founder, for everybody really on the team. So um, withoutea, just because it's a different scale, it's not massive. So we're not talking about like, I don't know, it's not a Volkswagen, it's not an IBM. We're not talking about hundreds of thousands of people, but there's thousands and thousands of people. And so all of a sudden like things are just covered. Wait, what do you mean? I don't need to take care of it? Security as well. We have a team for that. So like, like it's, it's just uh, it's, it's almost feels sometimes like a luxury. Um, okay, so that's, that's just fantastic, uh, to have. But what is, I would say great in particular about Altair in this environment is sometimes this can feel too divided. So like everybody's only doing the one little thing and that's not the culture of Altair at the same time. So maybe this is actually leading to the answer of the first half. Really, like, okay, what made this experience with Alteo so special and maybe even surprised me a little bit because I've been working for large organizations before is that everybody still feels this kind of like, um, feeling of ownership and there's a lot of passion and excitement. So when we joined I really thought like, what is going on here? Hundreds and thousands of people reaching out like, and being incredibly friendly. That's one thing and that's important. But the other thing, really, everybody wants to learn. Everybody wanted to be supportive and that is something I did not expect at that level and I think really did set Altair apart as a great new mothership. Really well.

Speaker A: They probably spent a lot of their time investing in the right people to, to have that kind of experience.

Speaker B: It probably has to be like that. Yes. Because I can't explain it otherwise.

Speaker A: Yeah, right. Well, I know we're sort of at time here, so I, I just want to thank you so much for your time for telling us all about Rapid Minor Altair, your personal journey of, you know, going through that experience, talking about Cambridge semantics. And it was just really, really great to learn more about this. Congratulations again to your little dance magic quadrant dance. There you go.

Speaker B: All righty. All righty. So now every probably now you get like, a dozen more questions like, who is this weird guy? And why did you even talk to him? Like, I'm sorry, folks, but, like, yeah,

Speaker A: no, this was great. I really appreciate you taking the time to chat with us. And, yeah, I encourage folks to follow Altair on LinkedIn. Go to the website Altair.com and we'll share the links for the Academy, which I think is also great, where you're investing in the community, telling them, how do you actually use the tools? Before saying, here are the tools. How do you use a hammer? Before you give me a hammer, what do I actually do with it? And which nail do I hit? Uh, so again, thank you so much for your time.

Speaker B: Well, thanks, Kate. Thanks to everyone joining us today. Thank you.

Speaker A: All right.