Snowflake Data Cloud Summit 2024 | Serge Gershkovich & Jaymin Patel | SqlDBM

Clips
News
More from Snowflake Data Cloud Summit 2024

Serge Gershkovich

Head of Product

SqIDBM

Jaymin Patel

Senior Manager, Data Platform and Engineering

SqlDBM

Unlocking AI potential with effective cloud-based data modeling

In today’s data-driven world, the key to unlocking the true potential of data lies in effective cloud-based data modeling.The ability to structure, share and collaborate on data models in real-time is revolutionizing how organizations manage and utilize data. Cloud-based solutions are at the forefront of this transformation, providing scalability, sharability and integration that traditional on-premises solutions cannot match, according to Serge Gershkovich (pictured, right), head of product at SqlDBM (We Comes Before Me Corp.).Baptist Health South Florida’s Jaymin Patel and SqlDBM’s Serge Gershkovich talk to theCUBE about cloud-based data modeling

play_circle_outline Exploring the Snowflake Data Cloud Summit: Insight into SqlDBM's Cloud-Based Data Modeling Solution and its Benefits

play_circle_outline Testimonials from Jaimin Patel on the importance of data modeling in healthcare

play_circle_outline Introducing TX: An All-In-One Relational Transformational Modeling Platform with Unbeatable Features and Availability

Info

Serge Gershkovich & Jaymin Patel | SqlDBM

Serge Gershkovich

Head of Product SqIDBM

Jaymin Patel

Senior Manager, Data Platform and Engineering SqlDBM

search

Rebecca Knight

>> Hello everyone, and welcome back to theCUBE's live coverage of the Snowflake Data Cloud Summit here in San Francisco, California. I'm your host, Rebecca Knight, sitting alongside my co-host, co-analyst, co-founder of theCUBE, Dave Vellante. Dave, the show is going great. Lots of great guests. And we're going to talk about one of your favorite topics, data modeling.

>> Yeah. Well, it's a topic that is really important because it's part of getting your data estate right, so you can apply AI to it.

Rebecca Knight

>> Exactly. Indeed. Well, I'd like to introduce our next guest. We have Serge Gershkovich. He is the head of product at SqlDBM. Welcome. Welcome back to theCUBE, I should say. CUBE alum.

Serge Gershkovich

>> Thank you you so much. Great to be here.

Rebecca Knight

>> And Jaimin Patel, Senior Manager, Data Platform and Engineering at SqlDBM. Welcome, Jaimin.

Serge Gershkovich

>> Thank you so much.

Rebecca Knight

>> So Serge, I'm going to start with you. Talk a little bit about SqlDBM, what you're building, what you're all about.

Serge Gershkovich

>> Of course. SqlDBM is a cloud-based data modeling solution, and data modeling has been around for as long as databases have been around. But the cloud and everything, this is why we're here today. Snowflake and scalability, shareability collaboration, this is what the cloud unlocks for just about any solution, any product. And data modeling is most useful when you get it out in front of the business, in front of your coworkers, in front of your users. It's something that starts with the data team, but is shareable and usable by pretty much anyone working on the database. So when you're able to just easily input somebody's email address and have them come into the same vision that you're designing and see it in real time, we feel like that's really what unlocks the true power of modeling. Because once you've designed it, you want everybody to be aligned on that same vision. And as soon as you take it out of your desktop and print it out or save it as a document, now it's instantly out of sync. So with cloud, you have shareability, you have integration, you have modeling in parallel, in real time. So it's really foundational and really powerful and transformative.

>> And the customers we've had on previously talking about SqlDBM, cloud-based, cloud-first, cloud-native, it's very simple to bring in, as you say, non-technical users. And my understanding is there's a little bit of art involved in data modeling, and so to the extent that you can align people on what that sort of work of art is, and you could do a lot with small teams, is sort of the value that you guys bring. Have you seen that and is that part of the appeal? I wonder if you could first of all describe your role and then we can get into the sort of discussion about how you're doing modeling.

Serge Gershkovich

>> Sure. So I'm senior manager of data platform and engineering. My role at Baptist is to get data from various source systems, about sixty-plus source systems, daily and make that available for our end users, reporting team, AIML teams and so forth. So as part of this, structure of data matters, if you don't have a structure to it, people are going to keep guessing and you're not going to have great adoption and usage of the data. So that's what SqlDBM helps us is to structure the data from raw to curated, which is modeling, having relationship within each of the functional areas for the financial and clinical domain. Easy view to that. So if you have a question, you can quickly find how you can relate the info and get insight to it and get that info to doctors, nurses, and so forth.

>> So you're a healthcare provider. And what's the scope of your responsibility? You're serving the entire organization or is it just sort of a narrow piece?

>> Yeah, I support complete corporate Baptist for all the data needs from platform management to provisioning them and making it available to the end users.

>> So what were you doing before SqlDBM and what led you here? Take us through that journey if you would.

>> So I think before SqlDBM there was some collaboration, but that's specific to a particular need or request. And when you are working on a specific project, you don't think about how it connects to all the other data that you have and what more insights can you get from it. So what SqlDBM helps with now is once we model it, we can collaborate with stakeholders, SMEs, and various other teams to understand is this model works for this project versus any other needs that we may have for the data. And collaborate and then finalize that model and make it available to be used by the end users.

Rebecca Knight

>> I am fascinated and excited about the potential of changing the healthcare industry, how it's delivered, how our experience as patients is changed. Can you just sort of bring this down to earth a little bit in terms of how even just our experience as someone who is going to Baptist Health to receive medical care would feel the difference?

>> Yeah, so let's get to the front end. So Baptist Health is a data-centric org. We have a PineApp, I'm not sure if you're familiar with it, but it's very popular. So where patient can go to the app, schedule appointments, look at their lab results and do more, message with the provider for any questions they may have. This all comes and is supported by the data that we collect in various systems and model it and make it available to that front app. So for patients, it's an easy access to their info in a protected manner.

>> So help us understand this. By doing the proper modeling, correct me if I'm wrong, but you have the semantics such that anywhere you apply that data, it's understood, it's coherent and consistent and that is a step that a lot of companies don't take. And then they'll plug their data into different BI tools. Billings means one thing in one place and another thing in another. Why is it that more companies don't start with a more disciplined data modeling?

Serge Gershkovich

>> That's a great question. I think even the term data modeling is a bit of a misnomer because it puts the emphasis on what the user interacts with, which is the database. But a data model is really nothing more than the business model in a readable, understandable set of semantics, just simplified. That we can talk all day about the nuances of what is a customer, what is considered an active customer, when was the last time they purchased or did they cancel, but if you can find a very simple set of semantics, both visual and something that translates into something a database understands, then you can just put it down in front of the paper. And now it's not one vision in my head, another interpretation of that in your head. We can look at that and say, "Oh yeah, we agree." Or, "Actually no, that is not the granularity of a product. It's actually this category is also involved." So it takes the ambiguity out of it and which point designing it as a database is trivial. And because data tends to just appear once you have a running enterprise, you have an order and the order comes in, you have a reservation and the reservation comes in with certain details. People just assume that they don't need to model because they have something to work with. Now I can transform it. Now I can apply some rules and business logic, put it in front of a dashboard, but is does not imply ought. And they have no guarantee that what they're looking at is representative of how their SMEs, how their field sales force is really looking at that same reservation or that same contract that they helped create. So it's really something that you can get by if you don't have it. I was just chatting with Jaimin before the interview and he mentioned that Baptist is actually a non-profit so efficiency is paramount. They have to do, like I said, more with less. And when you don't have a formal data model, you still have a data model. It's in somebody's head, it's in some system. It just means that anytime an analyst has a question, they're going to go and they're going to bug the lead architect or they're going to go bug the business users when they can really self-serve all of this information, if you can formalize and centralize it. It's just not an efficient way of anyone's time. Meetings, conversations, and even then kind of guessing at, "Okay, did he really explain it well? Did he understood what I explained?" And mature enterprises just, they don't survive without data modeling. The bigger you get, the more probability is that you have a central architect who manages the data model, who makes sure that data teams play within the confines. But when you're smaller, that might not be the case. It's like when you walk into a shopping center, if it's small enough, you might just say, "I'll just wander around. I'm pretty sure I'll find my way." But when it's Mall of America, you can spend three days and not without a map and not find what you're looking for. And this is the thing where these are two edges of the extreme and it's like a pot of water coming up to boil. You don't quite notice where that line is, like, "This is uncomfortable now." But it's also what you've been doing yesterday.

>> It's a great analogy. And when you think about what we heard yesterday in the keynotes, I don't know if you guys are in there, but all these announcements around Cortex AI and it's new, you're probably not playing around with it yet, or maybe you are. But unless you have a well-architected data model, you're not going to get as much value out of applying that. And to Rebecca's point, if you can apply that, then the service levels you're going to be able to give your patients, it's going to be much, much better. How do you think about applying AI? Where does that fit in the roadmap?

Serge Gershkovich

>> So we are doing a lot with AI, generating patient summary of their visit to quickly getting that details. But if you don't have the proper data model, the hundreds of tables, hundreds of data sets, how can you get to the info that matters to the patient or the physician to treat them and have a better outcome for the patient care? So this is where we try and holistically think about how this particular model for a subject area can help us to easily answer those questions. And let's say you are doing a model where you want to say bot is the bad utilization and how you can improvise that operationally, so that data helps them to quickly get to that instead of going for months without getting to that insights.

>> And so I was kind of going from the patient standpoint, but the doctors get a lot of benefit out of this because they don't have to sit there and type the whole time and the patient sits there. It's sort of awkward. And I presume the doctors must love this, right? I mean, it just takes away all that heavy lifting.

>> Absolutely. And then they can also do get further reports that can help, "Okay, this particular age category for a patient, this is what their diagnosis were. How can we use that for further research and improve the treatment plans and so forth, the other patient?" So there's a lot to it. Again, not going into much specifics of it, but having a proper structure on the data simplifies your use of data,

>> And it compresses the time at which the patient can actually take action. How many times have you been to the doctors and then the days later or they're getting a phone call, they leave a message, they get an email that gets buried. If you can actually say, "This is what you got to do right now, go do it." I mean, sometimes it's simple with a prescription, but something more complicated, that compresses the time to value.

Rebecca Knight

>> No, indeed. Indeed. So how closely are you working... I mean obviously you're working closely with your customers and partners, but in terms of gathering their feedback about how they're approaching generative AI and the potential that they see to transform their organizations? What are you seeing right now?

Serge Gershkovich

>> We're seeing that AI is... We can talk about kind of the role of AI within modeling, which is still a little bit unproven, but we're seeing the old paradigm of garbage in garbage out come to fruition like never before because AI is nothing more than a pattern recognition and sensing tool that will just kind of sort through what it has and give you the best possible prediction based on that. And when you have an unproperly modeled data set where the inferences that it can generate are just guaranteed to be incorrect, if the data is wrong, if the data's messy, if it has to do extra work where you have a customer called customer on one side, client on another, will it figure out that a customer is a client and they're the same thing? It's just adding more and more ambiguity to something that is already ambiguous by nature. And we're seeing a lot of interest come back into the data modeling space because all of the engineers training LLMs, the first thing they're going to ask for is a clean data set. Obviously with AI, we're going to save them the trouble of having to formally define of exactly how these things link together. That could be also inferred from a properly curated data model. But to go at it blind with untested, uncleaned, unproven data, and no foundation to even help you understand what the case is, that's just the recipe for disaster because LLMs, just like Snowflake, Compute, is not cheap. There's significant hardware running in there. So it's not just, "Oh, my query took a little longer." It's like, "I just blew through $10,000 of compute credit and it's giving me gobbledygook." So that's not going to fly. It's not going to fly.

Rebecca Knight

>> .

Serge Gershkovich

>> Yeah. Run it again. Exactly.

>> You guys have news, I'm looking at BusinessWire that you've unveiled. It says the first all-in-one relational transformational modeling platform, TX, you call it. What's the news? Fill us in.

Serge Gershkovich

>> Oh, great question. We're super excited about TX. So TX stands for transformation and it's the next logical step when you do modeling. You go from conceptual to logical, physical and it's still just describing what your data should look like. And then the next obvious question is, "Okay, well now go ahead and move it, transform it, put it in front of the business." This is just the outline so to speak. And this is the space where sooner or later, we just knew that we had to cross that threshold. Because models are great, they help streamline, they help guide the transformations. But everybody working around this convention floor today, anybody using Snowflake, they're using Snowflake to transform data. So why would we not make it easier for them? And if anything, we're not innovators. I'm surprised it took us this long because our customers just kept coming to us and saying, "Okay, well I've modeled it. I'm going to be using this tool to transform. I'm going to be using that tool to transform and I'll be hand-rolling my transformations. How can I get the data model and enforce it and ensure that it keeps up with what I'm now coding through logic?" And we said, "Well, that's it. Just do it all in one tool. We'll make it easy for you." And what we built is our existing foundational relational modeling in the same project as the transformation. So as soon as you create the data pipeline, it comes right back. Or it doesn't even come right back. It's already in the same project and viewable as part of your existing data model. So that allows you to work bottom up, meaning define the structure, validate it with the business before you've done the hard part, which is all the SQL logic and the business logic. And once they say, "Green light, go for it." You can just add the SQL or vice versa. The typical... The business comes and says, "Can you pull that data real quick?" Maybe you don't have time to go column by column, validate data columns. Just write the logic with a single button. We'll generate the model for you. Synchronize all of that and just again, bring it. They will be able to see the pipeline and the structure before it's even run and validated.

>> And when is this available?

Serge Gershkovich

>> It's already available.

>> Okay. So just shipping GA today essentially, or yesterday?

Serge Gershkovich

>> Yesterday was our big reveal and it's available for trials and demos. It's a SaaS tool. So that's the beauty of, again, running on Snowflake is we can build our beautiful front end and Snowflake will scale it. It'll just push down all that processing, the heavy lifting down to Snowflake and we don't really have to reinvent the wheel. Snowflake's already great at doing all of that.

>> So you haven't touched this yet?

Serge Gershkovich

>> No.

>> You're just getting started. So in your future potentially, what would it do for your organization?

Serge Gershkovich

>> I think, again, I've not been part of the review, but it's going to help us with how we are going to transform the data. So absolutely looking forward to the demo and-

>> That next step on the maturity curve is really what this is. My last question is what do you want to be able to say, let's call it 12 months from now, maybe 12 to 18 months from now that you're not able to today about your platform, your data estate?

Serge Gershkovich

>> I think we are going to continue optimizing and modeling new data sets and making it available for the end user and how we can use that data for AI/ML use cases, as well as supporting our patient care and for better outcome.

Rebecca Knight

>> Excellent. Well, Serge, Jaimin, thank you both so much for coming on theCUBE. A really fascinating conversation.

>> Thanks guys.

Serge Gershkovich

>> Yeah, it's always a pleasure. Thanks.

Rebecca Knight

>> I'm Rebecca Knight for Dave Vellante. Stay tuned for more of theCUBE's live coverage of the Snowflake Data Cloud Summit. You're watching theCUBE, the leader in enterprise tech news and analysis.

Serge Gershkovich & Jaymin Patel | SqlDBM

search

Rebecca Knight

>> Hello everyone, and welcome back to theCUBE's live coverage of the Snowflake Data Cloud Summit here in San Francisco, California. I'm your host, Rebecca Knight, sitting alongside my co-host, co-analyst, co-founder of theCUBE, Dave Vellante. Dave, the show is going great. Lots of great guests. And we're going to talk about one of your favorite topics, data modeling.

>> Yeah. Well, it's a topic that is really important because it's part of getting your data estate right, so you can apply AI to it.

Rebecca Knight

>> Exactly. Indeed. Well, I'd like to introduce our next guest. We have Serge Gershkovich. He is the head of product at SqlDBM. Welcome. Welcome back to theCUBE, I should say. CUBE alum.

Serge Gershkovich

>> Thank you you so much. Great to be here.

Rebecca Knight

>> And Jaimin Patel, Senior Manager, Data Platform and Engineering at SqlDBM. Welcome, Jaimin.

Serge Gershkovich

>> Thank you so much.

Rebecca Knight

>> So Serge, I'm going to start with you. Talk a little bit about SqlDBM, what you're building, what you're all about.

Serge Gershkovich

>> Of course. SqlDBM is a cloud-based data modeling solution, and data modeling has been around for as long as databases have been around. But the cloud and everything, this is why we're here today. Snowflake and scalability, shareability collaboration, this is what the cloud unlocks for just about any solution, any product. And data modeling is most useful when you get it out in front of the business, in front of your coworkers, in front of your users. It's something that starts with the data team, but is shareable and usable by pretty much anyone working on the database. So when you're able to just easily input somebody's email address and have them come into the same vision that you're designing and see it in real time, we feel like that's really what unlocks the true power of modeling. Because once you've designed it, you want everybody to be aligned on that same vision. And as soon as you take it out of your desktop and print it out or save it as a document, now it's instantly out of sync. So with cloud, you have shareability, you have integration, you have modeling in parallel, in real time. So it's really foundational and really powerful and transformative.

>> And the customers we've had on previously talking about SqlDBM, cloud-based, cloud-first, cloud-native, it's very simple to bring in, as you say, non-technical users. And my understanding is there's a little bit of art involved in data modeling, and so to the extent that you can align people on what that sort of work of art is, and you could do a lot with small teams, is sort of the value that you guys bring. Have you seen that and is that part of the appeal? I wonder if you could first of all describe your role and then we can get into the sort of discussion about how you're doing modeling.

Serge Gershkovich

>> Sure. So I'm senior manager of data platform and engineering. My role at Baptist is to get data from various source systems, about sixty-plus source systems, daily and make that available for our end users, reporting team, AIML teams and so forth. So as part of this, structure of data matters, if you don't have a structure to it, people are going to keep guessing and you're not going to have great adoption and usage of the data. So that's what SqlDBM helps us is to structure the data from raw to curated, which is modeling, having relationship within each of the functional areas for the financial and clinical domain. Easy view to that. So if you have a question, you can quickly find how you can relate the info and get insight to it and get that info to doctors, nurses, and so forth.

>> So you're a healthcare provider. And what's the scope of your responsibility? You're serving the entire organization or is it just sort of a narrow piece?

>> Yeah, I support complete corporate Baptist for all the data needs from platform management to provisioning them and making it available to the end users.

>> So what were you doing before SqlDBM and what led you here? Take us through that journey if you would.

>> So I think before SqlDBM there was some collaboration, but that's specific to a particular need or request. And when you are working on a specific project, you don't think about how it connects to all the other data that you have and what more insights can you get from it. So what SqlDBM helps with now is once we model it, we can collaborate with stakeholders, SMEs, and various other teams to understand is this model works for this project versus any other needs that we may have for the data. And collaborate and then finalize that model and make it available to be used by the end users.

Rebecca Knight

>> I am fascinated and excited about the potential of changing the healthcare industry, how it's delivered, how our experience as patients is changed. Can you just sort of bring this down to earth a little bit in terms of how even just our experience as someone who is going to Baptist Health to receive medical care would feel the difference?

>> Yeah, so let's get to the front end. So Baptist Health is a data-centric org. We have a PineApp, I'm not sure if you're familiar with it, but it's very popular. So where patient can go to the app, schedule appointments, look at their lab results and do more, message with the provider for any questions they may have. This all comes and is supported by the data that we collect in various systems and model it and make it available to that front app. So for patients, it's an easy access to their info in a protected manner.

>> So help us understand this. By doing the proper modeling, correct me if I'm wrong, but you have the semantics such that anywhere you apply that data, it's understood, it's coherent and consistent and that is a step that a lot of companies don't take. And then they'll plug their data into different BI tools. Billings means one thing in one place and another thing in another. Why is it that more companies don't start with a more disciplined data modeling?

Serge Gershkovich

>> That's a great question. I think even the term data modeling is a bit of a misnomer because it puts the emphasis on what the user interacts with, which is the database. But a data model is really nothing more than the business model in a readable, understandable set of semantics, just simplified. That we can talk all day about the nuances of what is a customer, what is considered an active customer, when was the last time they purchased or did they cancel, but if you can find a very simple set of semantics, both visual and something that translates into something a database understands, then you can just put it down in front of the paper. And now it's not one vision in my head, another interpretation of that in your head. We can look at that and say, "Oh yeah, we agree." Or, "Actually no, that is not the granularity of a product. It's actually this category is also involved." So it takes the ambiguity out of it and which point designing it as a database is trivial. And because data tends to just appear once you have a running enterprise, you have an order and the order comes in, you have a reservation and the reservation comes in with certain details. People just assume that they don't need to model because they have something to work with. Now I can transform it. Now I can apply some rules and business logic, put it in front of a dashboard, but is does not imply ought. And they have no guarantee that what they're looking at is representative of how their SMEs, how their field sales force is really looking at that same reservation or that same contract that they helped create. So it's really something that you can get by if you don't have it. I was just chatting with Jaimin before the interview and he mentioned that Baptist is actually a non-profit so efficiency is paramount. They have to do, like I said, more with less. And when you don't have a formal data model, you still have a data model. It's in somebody's head, it's in some system. It just means that anytime an analyst has a question, they're going to go and they're going to bug the lead architect or they're going to go bug the business users when they can really self-serve all of this information, if you can formalize and centralize it. It's just not an efficient way of anyone's time. Meetings, conversations, and even then kind of guessing at, "Okay, did he really explain it well? Did he understood what I explained?" And mature enterprises just, they don't survive without data modeling. The bigger you get, the more probability is that you have a central architect who manages the data model, who makes sure that data teams play within the confines. But when you're smaller, that might not be the case. It's like when you walk into a shopping center, if it's small enough, you might just say, "I'll just wander around. I'm pretty sure I'll find my way." But when it's Mall of America, you can spend three days and not without a map and not find what you're looking for. And this is the thing where these are two edges of the extreme and it's like a pot of water coming up to boil. You don't quite notice where that line is, like, "This is uncomfortable now." But it's also what you've been doing yesterday.

>> It's a great analogy. And when you think about what we heard yesterday in the keynotes, I don't know if you guys are in there, but all these announcements around Cortex AI and it's new, you're probably not playing around with it yet, or maybe you are. But unless you have a well-architected data model, you're not going to get as much value out of applying that. And to Rebecca's point, if you can apply that, then the service levels you're going to be able to give your patients, it's going to be much, much better. How do you think about applying AI? Where does that fit in the roadmap?

Serge Gershkovich

>> So we are doing a lot with AI, generating patient summary of their visit to quickly getting that details. But if you don't have the proper data model, the hundreds of tables, hundreds of data sets, how can you get to the info that matters to the patient or the physician to treat them and have a better outcome for the patient care? So this is where we try and holistically think about how this particular model for a subject area can help us to easily answer those questions. And let's say you are doing a model where you want to say bot is the bad utilization and how you can improvise that operationally, so that data helps them to quickly get to that instead of going for months without getting to that insights.

>> And so I was kind of going from the patient standpoint, but the doctors get a lot of benefit out of this because they don't have to sit there and type the whole time and the patient sits there. It's sort of awkward. And I presume the doctors must love this, right? I mean, it just takes away all that heavy lifting.

>> Absolutely. And then they can also do get further reports that can help, "Okay, this particular age category for a patient, this is what their diagnosis were. How can we use that for further research and improve the treatment plans and so forth, the other patient?" So there's a lot to it. Again, not going into much specifics of it, but having a proper structure on the data simplifies your use of data,

>> And it compresses the time at which the patient can actually take action. How many times have you been to the doctors and then the days later or they're getting a phone call, they leave a message, they get an email that gets buried. If you can actually say, "This is what you got to do right now, go do it." I mean, sometimes it's simple with a prescription, but something more complicated, that compresses the time to value.

Rebecca Knight

>> No, indeed. Indeed. So how closely are you working... I mean obviously you're working closely with your customers and partners, but in terms of gathering their feedback about how they're approaching generative AI and the potential that they see to transform their organizations? What are you seeing right now?

Serge Gershkovich

>> We're seeing that AI is... We can talk about kind of the role of AI within modeling, which is still a little bit unproven, but we're seeing the old paradigm of garbage in garbage out come to fruition like never before because AI is nothing more than a pattern recognition and sensing tool that will just kind of sort through what it has and give you the best possible prediction based on that. And when you have an unproperly modeled data set where the inferences that it can generate are just guaranteed to be incorrect, if the data is wrong, if the data's messy, if it has to do extra work where you have a customer called customer on one side, client on another, will it figure out that a customer is a client and they're the same thing? It's just adding more and more ambiguity to something that is already ambiguous by nature. And we're seeing a lot of interest come back into the data modeling space because all of the engineers training LLMs, the first thing they're going to ask for is a clean data set. Obviously with AI, we're going to save them the trouble of having to formally define of exactly how these things link together. That could be also inferred from a properly curated data model. But to go at it blind with untested, uncleaned, unproven data, and no foundation to even help you understand what the case is, that's just the recipe for disaster because LLMs, just like Snowflake, Compute, is not cheap. There's significant hardware running in there. So it's not just, "Oh, my query took a little longer." It's like, "I just blew through $10,000 of compute credit and it's giving me gobbledygook." So that's not going to fly. It's not going to fly.

Rebecca Knight

>> .

Serge Gershkovich

>> Yeah. Run it again. Exactly.

>> You guys have news, I'm looking at BusinessWire that you've unveiled. It says the first all-in-one relational transformational modeling platform, TX, you call it. What's the news? Fill us in.

Serge Gershkovich

>> Oh, great question. We're super excited about TX. So TX stands for transformation and it's the next logical step when you do modeling. You go from conceptual to logical, physical and it's still just describing what your data should look like. And then the next obvious question is, "Okay, well now go ahead and move it, transform it, put it in front of the business." This is just the outline so to speak. And this is the space where sooner or later, we just knew that we had to cross that threshold. Because models are great, they help streamline, they help guide the transformations. But everybody working around this convention floor today, anybody using Snowflake, they're using Snowflake to transform data. So why would we not make it easier for them? And if anything, we're not innovators. I'm surprised it took us this long because our customers just kept coming to us and saying, "Okay, well I've modeled it. I'm going to be using this tool to transform. I'm going to be using that tool to transform and I'll be hand-rolling my transformations. How can I get the data model and enforce it and ensure that it keeps up with what I'm now coding through logic?" And we said, "Well, that's it. Just do it all in one tool. We'll make it easy for you." And what we built is our existing foundational relational modeling in the same project as the transformation. So as soon as you create the data pipeline, it comes right back. Or it doesn't even come right back. It's already in the same project and viewable as part of your existing data model. So that allows you to work bottom up, meaning define the structure, validate it with the business before you've done the hard part, which is all the SQL logic and the business logic. And once they say, "Green light, go for it." You can just add the SQL or vice versa. The typical... The business comes and says, "Can you pull that data real quick?" Maybe you don't have time to go column by column, validate data columns. Just write the logic with a single button. We'll generate the model for you. Synchronize all of that and just again, bring it. They will be able to see the pipeline and the structure before it's even run and validated.

>> And when is this available?

Serge Gershkovich

>> It's already available.

>> Okay. So just shipping GA today essentially, or yesterday?

Serge Gershkovich

>> Yesterday was our big reveal and it's available for trials and demos. It's a SaaS tool. So that's the beauty of, again, running on Snowflake is we can build our beautiful front end and Snowflake will scale it. It'll just push down all that processing, the heavy lifting down to Snowflake and we don't really have to reinvent the wheel. Snowflake's already great at doing all of that.

>> So you haven't touched this yet?

Serge Gershkovich

>> No.

>> You're just getting started. So in your future potentially, what would it do for your organization?

Serge Gershkovich

>> I think, again, I've not been part of the review, but it's going to help us with how we are going to transform the data. So absolutely looking forward to the demo and-

>> That next step on the maturity curve is really what this is. My last question is what do you want to be able to say, let's call it 12 months from now, maybe 12 to 18 months from now that you're not able to today about your platform, your data estate?

Serge Gershkovich

>> I think we are going to continue optimizing and modeling new data sets and making it available for the end user and how we can use that data for AI/ML use cases, as well as supporting our patient care and for better outcome.

Rebecca Knight

>> Excellent. Well, Serge, Jaimin, thank you both so much for coming on theCUBE. A really fascinating conversation.

>> Thanks guys.

Serge Gershkovich

>> Yeah, it's always a pleasure. Thanks.

Rebecca Knight

>> I'm Rebecca Knight for Dave Vellante. Stay tuned for more of theCUBE's live coverage of the Snowflake Data Cloud Summit. You're watching theCUBE, the leader in enterprise tech news and analysis.