Energy Bytes

Igor Kuvaev, CTO of ROGII Inc. discuss how cutting-edge technology is transforming the oil and gas sector. Igor shares his journey from geologist at ConocoPhillips to leading innovation at ROGII, where the company is using AI to unlock hidden profits and revolutionize drilling operations. The conversation covers the challenges of outdated technology, the essential role of geosteering, and ROGII’s approach to real-time data integration and edge computing, setting new benchmarks for the industry.

Join the conversation shaping the future of energy.
Collide is the community where oil & gas professionals connect, share insights, and solve real-world problems together. No noise. No fluff. Just the discussions that move our industry forward.
Apply today at collide.io

00:00 - Intro
01:15 - Igor's Background
02:38 - Role of a Geologist
06:54 - Evolution of Geosteering
14:01 - StarSteer Technology Stack
19:10 - Real-Time Operations at StarSteer
24:46 - Cloud Agnostic Solutions
29:01 - Data Sharing and Python Integration
31:05 - Python SDK Overview
34:30 - User Feedback Importance
38:55 - Benefits of Python
43:18 - Exciting Future Developments
43:36 - Edge Computing Innovations
48:49 - Future of Self-Driving Cars
50:53 - Lightning Round Q&A
56:15 - Advice for Energy Tech Startups
58:05 - Solving Problems in Tech
59:02 - Favorite Game Discussion

https://twitter.com/collide_ai
https://www.tiktok.com/@collide.io
https://www.facebook.com/collide.io
https://www.instagram.com/collide.io
https://www.youtube.com/@collide_io
https://bsky.app/profile/collide-ai.bsky.social
https://www.linkedin.com/company/collideai

What is Energy Bytes?

Welcome to Energy Bytes with John Kalfayan and Bobby Neelon. Your essential guide to the intersection of data and energy. This podcast dives deep into the world of energy, shedding light on how data, AI, and technology are revolutionizing this sector. Each episode equips listeners with insights into the most efficient tools and resources, paving the way for a data-driven future in energy. From technical nuances to broader industry trends, Energy Bytes offers an unparalleled perspective on the evolution of the energy industry. Join us as we decode the algorithms of energy, one byte at a time.

Welcome back everybody to another episode of Energy Bites here with Brad Dad, me John Calfan. Got before we get into anything, I just want everybody to uh stop real quick, go give us a a review. Hopefully it's five stars, but give us some feedback. Tell us tell us what you like, what you don't like, what you want to hear. Um also please, you know, subscribe and share um with this so we can keep keep going and get some more more guests and stuff. So Um here today again with my trusty co-host Bobby Neelon. How's it going? How's it going, Bobby? Doing well. Excited we got Igor Kovyev with with Rogi. Rogi. Rogi. Um and now it's just fun for me cuz I think first time I've other than John got a former co-worker on the podcast. Igor and I used to work on the Eagleford uh Eagleford development team and at ConocoPhillips. So excited to have him on and you know, kind of hear his story cuz all I knew was he went from Conoco and all of a sudden boom he's you know, CTO of Rogi and definitely want to unpack that story and how he got here. So welcome. Thank you. Thanks for inviting. Looking forward to this conversation. Yeah, thanks for coming. But yeah, I mean I guess I mean I got to give a little bit of intro, but definitely not not enough. You want to just kind of tell us about yourself, you know, your background, where you're from, how you got into oil and gas, and all the you know Yeah, yeah, absolutely. Yeah, so my background I'm I'm a geologist. So I graduated as a geologist. I've been working as a geophysicist/geologist my entire career. And pretty much my career can be split into like two major parts. First one working for ConocoPhillips in various different locations. Uh as a geologist of course. And uh one of those locations was obviously United States here in Texas. So as Bobby said, I was working on on Eagleford play when it was just starting, you know, doing bunch of experiments there, drilling those first horizontal wells, you know, uh checking what kind of production is there, you know, and doing, you know, a lot of interesting stuff. And then I with my co-founder we were working actually together in ConocoPhillips. So, we quickly realized that there is a amazing opportunity for further geosteering improvement because back then the technologies were not developed, let's put it this way. So, we quit Conoco and founded our own startup software startup. Awesome. Yeah. Um Yeah, so let's kind of make it into that. I mean, you're talking about like the technology wasn't developed. I'm trying to think back we had the operations geologists, right, at Conoco and like they had to be in the office, right, and they had to be on the like these they call them the Linux blades, right? I mean, so what kind of what was it like then and then, you know, what opportunity did you see and how did you kind of pivot? Like, was there new technology available that they weren't that wasn't being utilized? Before before you answer that real quick, just for the more tech-focused people that may not know as much about the energy space, just give a quick kind of overview background of what a geologist does at an operator. And and maybe get into what geosteering is. Yeah. Yeah. Yeah, kind of go from geology, like what the geologists typically do, where they're involved, and then how that ties into geosteering specifically. Yeah. Yeah, absolutely. So, I think speaking about geology, uh I will be specifically focusing on geology of unconventional plays because this is what we do the most and it's substantially different from what geologist is doing for conventional plays, right? And for those who are not generally familiar, unconventional is what is being developed in United States the most because all the conventional plays, like old old school plays, they've been developed, so the oil has been produced, so we switched on conventional development. And this is what we do. So, as a geologist, basically what you do in unconventional space, you plan wells, so you identify where uh wells are going to be drilled, and those wells are always horizontal with very minor exceptions, but let's say, you know, always horizontal wells. So, you plan those wells, you plan where the well's going to be drilled like on the map view, where you're going to be placed those wells, where those wells will be placed uh in the reservoir because the reservoir is, you know, fairly thick formation, you know, from maybe, you know, 50 ft to, you know, 1,000 ft. So, you need to know specifically where the well is going to be drilled, where the well is going to be placed. And then, as the well is being drilled, you as a geologist, so this is actually what uh we do the most, you as a geologist, you control how exactly the well is placed. So, you are in a constant communication with uh drillers, direction drillers, drilling engineers to make sure the well is actually being drilled where it needs to be drilled. Yeah. Now, I think that's a a very good explanation. I think a lot of people, like I said, just don't understand, right? Uh, it's one of, to me, one of the, like, engineering marvels that is so kind of hidden in the energy space; is the fact that we can start with, you know, a whole whatever the conductor casing is, uh, you know, 30, 20, 30 in around at surface and end up 2 mi down and 3 mi away from that one spot with a pretty tight degree of, uh, error, so to speak. Um you know, it's pretty fascinating. And And we just take it for granted that we we really do. Right. Like and, you know, drilling is not easy by any means vertically, no less, turning out and going, you know, horizontally and stuff. So, I just wanted to highlight that because I don't think a lot of people really appreciate how, you know, you're landing a 5-in drill bit within, you know, within a few feet of where your plan or design was. Yeah. I mean, and Rogi's a great intersection of like that physical technology along with now just, you know, integrating with, you know, great softwares. Yeah. Yeah. And uh I think uh what uh what is also important to understand that because we are we're dealing with subsurface, uh uh it's uh we're very limited with the amount of data we receive, right? Because radio signal doesn't work over there. Uh there is no GPS. The there are like literally nothing is available to us. Uh the data transmission rate between our downhole tools, it's you know, 1 bit per second, not 1 byte, 1 bit per second, right? Wow. So, we we literally need to make a multi-million dollar deci--sions based on absolute, you know, absence of information. We we don't have much. Uh so, uh uh and we we can talk about how it started, what what's going on right now, but yeah, it's it's it's a super super exciting field where where you need to react quickly based on nothing. Yep. Yeah. Yeah, so yeah, I mean, maybe you like just in general, not even like say from Conoco's standpoint, but just like what did geosteering look like when you started say in the unconventionals, and then how's it progressed? I mean, now I mean, you guys pretty much own most of the market share as far as I understand, you know, as far as geosteering in North America. A lot of I mean, like so, how's it how has that technology kind of transformed over the last 10 15 years? And yeah. Just, you know, what did you guys do that just accelerated it? Yeah. Yeah, absolutely. So, we um uh I mean, back back then before before we started, uh the geosteering was quite I would say quite an interesting science, and it wasn't actually even I would say formed or created as science. Everybody was doing their own thing. Right. Uh if you go to a conventional play, uh the geosteering would be some uh uh I would say uh pretty much dependent, like field dependent. Each operator, even each uh oil field would do their own thing, and it would not be necessarily, you know, clear the way they do that. I started drilling my first horizontal wells back, uh you know, in 2000, maybe before 2010. And it was always uh very unclear procedure, and it was on conventional fields. Now, when I moved to United States, I saw the the big and dramatic difference between pretty much United States and all other countries is uh amount of downhole tools uh United States operator use. It's it's literally nothing. It's gamma ray only, which is the simplest and cheapest downhole tool you can find. And um what shocked me here is that those uh like $10 million decisions, I mean, back then the drilling was expensive. Yeah. Uh is is is is is uh is made based on that uh cheapest downhole tool on Excel spreadsheet. Right. So, that that how it was. I mean, yes, back then we as geologists we had some software, but those software were not far away from just, you know, doing basic uh you know, math calculations in in Excel. And um this is what uh you know, totally shocked me, and then my co-founder were like, "No way. I mean, it it must be a better way. It must be, you know, better way to make it uh easier, and uh reduce amount of errors." And bunch of errors were done because uh people were not taken into account all available data, not necessarily real-time data, but the data that was available to the geologists at the time of the drilling. Right. Uh and those errors could not could not be made. I mean, we could we could do something. We could do much much better. And uh yes, so then uh we we started the company. Uh our main idea, I would say, our first approach to the uh to geophysicists, to geologists was, "Hey, we should be able to uh integrate all available data, uh and by doing that uh drill better wells." And of course, you know, switch you from, you know, Excel-based software to something that, you know, an average person can use and enjoy. Yeah. So, those two components, maybe the second one more than the first one, was uh something that picked up, you know, very very quickly. I mean, very quickly we pretty much to cover the market. What uh what additional data were y'all kind of adding? Is that just like pilot test well logs and you know, kind of being able to interpolate between different actual uh data points or seismic or is it kind of all of the above? So, uh as far as uh just science data before you drill or when you drill, you actually have a bunch of data available. So, yeah, those could be uh nearby pilot holes and uh uh I would say 100%of unconventional plays, they uh happen somewhere next to the existing conventional fields. So, they're very heavily drilled. So, actually we do know a lot about of about geology already, right? So, we have a lot of that information. And we uh so, we have those pilot holes, we have structure maps, we have uh seismic data, we have uh mm nearby drilled horizontal wells. And uh the more we drill, the more information we get. So, the more data we can integrate. So, uh that was uh uh the first uh uh kind of idea that we brought to the market that was very well received. Now, uh going forward, uh we definitely should and we already can integrate some uh real-time drilling data, right? Because we have much more drilling data in real time versus the downhole uh gamma ray data. So, that data can also be integrated. But, the first like the core I would say the core uh thing that really transformed the industry and I would even say uh moved uh just even from like strange Excel spreadsheet into a real science, real job was that, you know, massive just science data integration. Yeah, and what there's I think that's been one of the kind of lowest hanging fruits in the oil field across the the operation is just we have all this data, right? But, it's siloed, it's in different softwares. It there hasn't been a lot of at least until last 5 10 years, right? Hasn't been a lot of effort it seems like in the energy space to try and you know, it's like we capture all this stuff, but then using it is much more difficult. Uh, and that's why, in my opinion, we've seen all this kind of energy tech boom over the last 10 15 years in the in the space, because there is all this opportunity, especially from That's what I love about it is you've got people that, you know, you were a geologist at an operator and you saw that, "Hey, we have all this stuff, but there's no solution to do it how it would make the most sense," and so you go and start your company, and now you guys are kind of the the dominant player in the space. It's It's It's awesome. I love I love that. And I would I would even add uh uh using it in real time. It's It's another level of complexity because technically even before us, I mean, can you take a look at your structure map on seismic? Yes, you can, you know, as soon as you are in the office you can take a look at it. But the rig doesn't care, you know, the rig you know, okay, some companies are out there they not in US, but I've seen those like, oh, yeah, let's stop the drilling for the night and figure it out in the morning. Uh no. No. It It's not It's not I mean, you know, rig to rig is what, 7 days right now or, you know, 10 days? Uh you you cannot do that, right? The rig has to drill. So, you need to have access to all of that data in real time by yourself. You don't have a luxury of, you know, stopping by and scheduling a meeting with your, you know, geophysicist uh you know, daytime. Take a week, figure it out, yeah. Now, that's another big kind of logistical like to me the oil field is so difficult because it is this giant logistics game, right? Everything costs money. Everything is tied to this ever-changing commodity price ultimately. And so, you know, yeah, there are much we could do these things much easier and potentially better if we had more time, but then you've got the cost perspective of that, right? And so, it's it's always this delicate kind of balance of of time versus money, but do y'all let's get into kind of y'all's kind of tech stack and and that sort of thing. Kind of how is that cuz y'all are I assume cloud-based in some aspects. Um but y'all also have hardware on site, like a any kind of edge devices yet, or is that kind of more future? Kind of what's what's your current kind of stack look like without You don't have to give who or if you don't want to, but um just what does that kind of look like? So, at the moment, uh a majority of uh of things we're doing is software based. We have edge devices as well. I can talk about it a bit later, but it's it's a cloud-based solution that is hosted on Amazon. Uh and uh I would say 2/3, maybe 3/4 already of our North American customers, they're all on cloud. So, they store all of their well data um uh all of their geo science data on our cloud. In fact, at the moment we've got roughly 1 million wells stored on our cloud, uh and I would say on average 20,000 wells are being added every month by the operators. It's not we, it's the operators. It's it's their own secured cloud, so uh they store the data, and uh they uh access this data either in real time or later on for analytical purposes, whatever they want to do with the data. Um now, uh we, even though, you know, even though we're cloud-based, there is always, you know, a front end part, right? You know, because this is what, you know, this is the most important to the users, you know, what what's your front end. And uh for the front end, we have um uh desktop application that is uh connected seamlessly connected to the cloud, uh uh and uh allows users to basically manipulate the data on the cloud, which is our Star Steer, the GeoSteering software. And we also have a bunch of uh uh web-based applications for monitoring. So, those ones are typically used uh yeah, either just by geologists. In fact, we have like iPhone, Android apps, so they can just quickly see what's going on on the rig. Or uh they are used by direction drillers or even company men. So, at the rig site, they can actually see, you know, what what's going on. So, this is actually the main uh one of the main advantages of the cloud. Uh it's kind of enabling these true real-time operations. This is something you cannot do without the cloud. It has to Yes, yes, it has to be on the cloud. It has to be updated in real-time. That's That's absolutely important for the modern drilling. Um we do have a piece for the data. So, most of the cases we rely on the data transmission from the rig that is uh uh that is established by um EDR or rig providers. But, for some cases, I would say mostly for outside of United States cases, we do have our own edge device that we put on the rig that actually collects all the data on the rig and then transmits that data again to the cloud so everybody can use it. Is that all with XML? Yeah, with XML. So, okay, the transmission part with XML uh as as many things in oil and actually many other industries is is kind of antiquated protocol, but uh It's standard. It's been there and at least it's standardized. It works. Okay, it works. That's That's the most important. Uh we as far as the data transmission, we have our own super compressed protocol because this is the bottleneck, the data transmission part from the rig to to to the cloud, right? But, once it once it gets to the cloud, yes, it's XML. Okay. How do y'all Let me ask cuz I know this is a ever evolving kind of question in the kind of field IoT space, but what are y'all mainly using for comms for those? Are y'all using satellite, cellular, or like Starlink? Have you messed with any of the some of that stuff yet? We when we put our own edge device, we rely on existing internet connection that, you know, typically rig operator is doing. We did We did some experiments with or maybe some some tests with things like Starlink. It's It's It's an interesting technology. The biggest problem of Starlink is that it has to be always at the same place. It's by by contract. Okay. So, it's not really very well suited for the moving devices, right? So, I don't know if anybody is using it for like rig operations. Maybe they are, maybe they have some special special things with them. we use it in the field, but again, I think it's more static site right now. Or even like But there's a lot of like not even necessarily Starlink, there's a lot of low orbit satellite stuff that people are doing now, too. But But again, it sounds like especially North America, you've got internet connection on most of the rigs that you can utilize. So, it's always nice that someone else can bring it and you don't have to worry about it yourself. I know it's generally most of the places these days, very remote places, everybody has internet. I mean, it's cheap. Satellite internet is cheap. So, luckily, the connection is not is not a big deal these days. Yeah. Talk to us a little bit about kind of the the architecture process and the thought process that you'll have gone through because, you know, so, my like I was saying earlier, you know, my background on the frac side, we have a ton of data, none of it is standardized. So, at least you'll have that going for you on the on the drilling side, but we also have the same problems, right? You've got a ton of data, you've got real-time operations, it's expensive, people back in the office want to be able to see the data. But you also want to try and leverage as much of that data both real-time and existing to make on-the-fly decisions. And so, kind of how did y'all kind of approach that because it's not it's not an easy thing when you're dealing with, you know, real-time data, data or connectivity intermittency, or the fact that it can just drop out. Um you know, there's all those kind of nuances of the oil field um that you have to think about. So, kind of just take us through, you know, high level how y'all went about kind of approaching that or Mhm. looking at some of those problems. So, once we decided to switch So, initially, we developed StarSteer as a as a desktop application. So, that was the logical thing to do because again, you know, you switch from Excel to I mean, you do desktop better, yeah. don't want to go from horses to spaceship right away, right? You need You need a, you know, some intermediary point. So, uh we once we designed StarSteer, we started to think, "Okay, you know, obviously, I mean, everybody understood that, you know, for real-time operations, cloud is essential. What do we do about the cloud?" And uh as far as uh uh typical real-time data, which would be your survey data, it's it's a trajectory of the wellbore, so this is this gives you the location of the of the data. And real-time gamma logs. So, those two, let's call it objects, are fairly small in size. So, as far as architecture, I wouldn't say it's not a big deal, but it was fairly easy to to make it real-time because it's small amount of data. Uh it can be downloaded from any location with even very bad internet. And of course, you know, we enabled compression, so it made that data really small. When it got really interesting is is when we decided to add some additional geoscience data. And geoscientists, they like huge data. Not necessarily, you know, uh it's uh it's it's huge in value sometimes, but definitely it's huge in time in in in volume. So, those would be structure maps and seismic data. Uh and nearby wells. So, you know, sometimes, you know, think of like Eagle Ford Permian, I mean, there are like thousand wells next to you, right? And you open a project as a geologist, you would like to see like few hundreds of them. You would like to see your seismic data, you would like to see your grid. I mean, that entire project could easily be, you know, over a gigabyte in size. And we need to bring it to the user that is located somewhere in Permian Basin, right? With not very good internet. And it needs to be in real-time, and it needs to be fast. And the user, you know, geologist doesn't want to hear about, "Hey, your internet is limited." You know, it's it's now I I'm the provider of the solution, so now it's my problem, you know, even though it's not my internet. So, this is where we develop um a lot of um fairly sophisticated tools uh on uh uh on data compression and uh on um like buffering the data. So, let's say you download the data once to your computer and then you go to the field and uh it it checks whether the data has been modified and if it's it hasn't been modified, it won't download the data anymore. Okay. So, uh additionally, the hardest part for us was the 3D seismic because believe it or not, some 3D seismic volumes could be as big as uh you know, 200 uh GB. Yeah. Uh and that's that's onshore US. So, this is where, you know, the wells are being drilled. And of course, again, the users, they would like to have the access to the data right away and uh you know, even with like almost no internet. Uh you know, so uh uh they we uh luckily, there is a very good uh seismic compression technology available uh which we haven't developed. It's uh company named Blueware out of Norway initially. They developed it. And they made it open source and uh we collaborated with them. Speaking about successful collaborations, right? So, you have to pay a giant fee to be a part of a big group to collaborate with them or was it just a native hey, this makes a lot of sense and we should do this? No, they they just uh they just uh provided it to us and uh uh it's it's an amazing technology. Uh we build it uh in into into our solutions and now I mean, believe it or not, you know, even 200 GB uh seismic data, you can uh view it in real time, you know, while while, you know, being on site. So, it's it's it's it's an amazing thing. Yeah, yeah. And it's one of those things, right? Innovation is is making something easier, right? It it was possible before, right? You could you could call your, you know, geophysicist and hey, send me a screenshot. Sure. But the innovation is hey, now I I can, you know, slice through my data located in the middle of nowhere. Sorry to poor men, guys. And and have access to the data in real time, right? Yeah. Yeah, you can see it when you're on vacation on your on Android or Yeah, exactly. or Apple. Mhm. Um No, that's Maybe yeah, cuz I was going to ask like if you if it start how it started it always start cloud native or is that kind of the the logical progression? So, um

Yeah, so I mean so how How do you guys arrive on AWS is basically was the main main player at the time and you know, I'm assuming it's it's been a been a good decision. Um How, I mean, let's say even from where you started with AWS to now have you has had to re-architect, or have you kind of been able to carry on with the same, kind of using the same tools? Cuz, I mean, obviously, like you know, Amazon RDS is different than it was, you know, 8 years ago, or I mean, depending on the different tools you're using. Have you guys had to kind of pivot or has it just been changing as you've added new features? We we changed a lot, but the logic we had from the very beginning was that we shouldn't rely on one single cloud data provider. Okay. So, and it's it's a fundamentally different development from what I've seen a lot in the industry is when you're you heavily rely on specific microservices provided only by one single cloud provider. Again, it's not it's not something is better or something is worse, but for sure for us it takes much more development efforts because we are literally cloud agnostic. We can switch to Okay. uh Microsoft next day. Uh but on a flip side the good thing is uh we see uh that many operators outside of the United States they prefer to have cloud internally. Again, we can debate whether it's good or bad. Probably it's bad, but um uh that helps us a lot because being cloud neutral Right. we can go ahead and install uh the our solution completely, you know, within their system. So, they have a private cloud, then you can go put it on there's and it just probably needs Linux VMs and whatever. Yeah. They they have everything internally, and that internal part is not connected to outside internet at all, and there are a bunch of other requirements that you have to follow, and um I don't want to speak for for others, but I believe for those ones who build something, you know, purely based something based on Azure or AWS, it either impossible or will take years to to convert. So, refactor. Yeah. Yeah, you basically have to build it from the scratch. Yeah. Yeah, so you guys have been, you know, fairly focused on being diagnosed so you can be portable and you know Yes. Yes, that that was from the day one. Yes. No, I I mean, I think that's a fantastic kind of tidbit piece of knowledge that anybody that's out there in the tech space currently or looking to develop something, if you have any intention of I'll turn to the camera and be dramatic about it. But if you have any intention of selling software, especially cloud-based software, internationally, this will be a very big thing. Um right? Cuz a lot of a lot of people don't realize this, but a lot of the NOCs and national oil companies have like data residency policies, right? Where the data physically cannot reside on a server outside of the country even, right? And so you have to have options to be able to install your software on their on-prem private cloud or you don't even get considered, right? Is And that's extremely serious. I mean, uh you can be put in jail in those countries for trying to bring data outside of the country and it has happened. Yeah. No, I had a I had a a guy that I got my master's with who worked at Baker in mostly Latin America and he uh he told me one time he had to go there was an offshore job, I think in Brazil. Um, he had to go out to the rig, pick up the laptop, come back on shore, drive the laptop to the office, hand the laptop over to the operator, and then, like, wait in the lobby for 4 hours. And then they came back and had completely erased the entire laptop, and here you go. Right? Like, it's a very real thing, you know? And so, that's just, again, I want to bring that up because, in my past experience on the[] when I was doing edge stuff, that was a much bigger thing that I didn't even consider that pretty much every national oil company kind of has in place. So that's really interesting that y'all built it completely agnostic. It's really smart, actually. Yeah. I mean, and so one thing maybe to pivot in a different direction that I mean that's about keeping the data secure and that's extremely important, but you also have some other kind of integrations and kind of data sharing capabilities through your Roam E Cloud and everything. So I mean cuz one thing I want to get to get into is you also you have a Python module where people can write Python within Star Steer, right? Am I Is that correct? So I mean please talk about You talk about that innovation and then ultimately I'd like to kind of get into data sharing cuz I know you can share with Snowflake and different people's clouds and Spotfire, which is a big deal for a lot of people. So now but maybe we want to start cuz I always found that very interesting and I think John and I always talked cuz I know there's you know, plus some crypto stuff where he could write in his own little language, you know, in some of those crypto bots. But like say like you geologists who know Python can come into your platform and write their own functions and do their visualizations custom within your platform as well, right? That's that's actually that's actually a very good question because initially when we developed So yes, we have bunch of interesting integrations and we started with basically, you know, having Python within Star Steer, so a person can code. And when we when we started it, I thought that's it's pretty much feature for myself so I can enjoy coding Python while, you know, you know, on my long flights or whatever. But it picked up very well. So amazingly, geologists who are not coders by, you know, by education, they they picked it up. We see a lot of Python coding in different companies. In in my previous operator where we worked together in many many other companies, they sometimes they even have dedicated group of people who code scripts under StarSteer. And it's always very interesting customizations here, interesting scripts. Mostly are all related to the field of like data integration, maybe sometimes data cleaning, data data analytics. Because they have to think about those projects like thousands of wells, right? There is so much, you know, information you can get out of it if you just analyze this data as a whole, not like well by well basis. Yeah. So, then we because we have our solution on the cloud, we uh develop what is called Python SDK. So, this is also Python, but it allows any other companies to pull and push data from from our cloud and to our cloud. And that was also very well adopted, but actually the group of people who are using it is slightly different. It's typically either, uh, like true data not geologist anymore, like true data analysts within operators, or different service companies that would like to, say, pull the data, uh, and make the data available to direction drillers, make data available for, you know, to completion engineers, to whoever needs geoscience and and drilling data in real time. Because uh the uh the beauty of our real-time cloud is that the data gets to our our cloud first. So, if they would like to get the data, and uh uh it's uh the easiest or, you know, the the easiest ways is just to to go to go to our cloud, and it's always latest and greatest up-to-date with the interpretation available. So, they can they can take either the raw data or the interpretation, and they can use for whatever they want to. Okay. No, very cool. Um And so, I'm I'm assuming like that Python SDK would be kind of originally how people were able to, you know, do ETL like pull data into their data warehouses or into their Yeah, within their firewalls or whatever. Um But I mean but now is it is it I forget. Uh I've seen some of your brochures but you have like Rogi Cloud or the different ways of like sharing cuz I mean HVR you can share it to a SQL server, you can share it to Snowflake or BigQuery, you know, whatever people are using. Is that through the Python SDK or do you all actually have native connectors and or that stuff as well? Uh it's it's kind of through Python SDK but it's developed internally. So basically we have our internal um development code Solo Connect which is which is which is a small application that pulls the data out of our cloud and stores it into either SQL or Snowflake or whatever the database operator prefers. And uh yeah, the logic is is is probably like you just described is now you have it all in SQL in, you know, in the way you prefer it to have with all specific data, all claims, you know, the quotations, raw data all together. So now we can put together our own dashboards. And of course, you know, all operators they sort of compete with each other for the uh the the prize of the best dashboard available. And

Uh they all develop their own dashboards. Uh sometimes uh we provide them some templates. Okay, guys, this is what you can do and uh we constantly in communication with them, see, you know, maybe you'd like, you know, this specific functionality so we can, you know, add some more into those um SQL SQL tables and you can, you can pull more data or you can make it, you know, more uh uh um more attractive to your current operations, more useful, because yeah, all dashboards are beautiful, but the question is which of those are useful, right? Uh yeah, yeah, exactly. Just like the whole models are Yeah, yeah. uh wrong but some are useful. But um So I mean I think looking at all those whether it's like the original Python models or Python SDK or you're obviously iterating on like dashboards or you know, how people want to see the data in SQL. How's it help kind of shorten feedback loops for your software development or I mean have you got a lot of like ideas for the you know actual platform from you know seeing what people are using the data outside of it and Absolutely. Absolutely. We we strongly believe in the constant communication with the users. And this is probably where users feedback is so much more valuable than like managers feedback, right? Cuz the typical question that we try to ask users, please don't don't talk about greatness of Star Steer. And if you want to, please go ahead, but the most important is please talk about your pain. Tell us what what you dislike. Tell us what concerns you or maybe about some things that you would like to do that you cannot do at the moment. And and show me as a separate question to what you just said, show me, you know, maybe your dashboards or whatever you can show me so I can better understand your workflow and your challenges. Because uh when when we see all of those different approaches and companies are still different. I mean there is general workflow that everybody is doing, but as far as like truly innovative thing, everybody is doing their own thing. So, show me what you're doing and when you see all of those things all all over the world, you can you can better understand what what users actually want, how how you can make their life easier. If you see a certain pattern showing up multiple times, you know, you're like this is something we need to prioritize. There's definitely a demand for it and I was going to ask you what cuz you're talking about databases with the connector. What are Are you seeing more adoption of like Snowflake from MySQL or is there kind of a Is SQL still the kind of common database that people are still serving to? SQL is the by far the most common, but definitely users or they they they tend to store more and more data. I mean we we all see that trend. It's more and more data, especially going into the real-time drilling data. I mean we're you know, gigabytes per well of data. And now you want to do a fast analytics based on like hundreds of those wells, right? So, we're talking about huge amount of data and ability to quickly process that huge amount of data. So, I think this is uh we here we're going into even like no sequel type of space, real-time databases and it's it's something where the technology is being very quickly developed over the last, you know, few years because just sequel cannot do it anymore, right? So, Snowflake is a good one, I think. Many companies are uh switching to Snowflake. Um there are, you know, others no no sequel database, I would say in this that are in the stage of experiment right now. But, I guess the good thing about us is even though we have a lot of data, we don't have as much data as many other industries, like, you know, automotive or just manufacturing or IoT. So, there are databases that built for those guys. We don't need to reinvent the wheel. We can just, you know, steal ideas from them. No, absolutely. Yeah, and I and I guess there's things too like that's what you're seeing on the operator side, but I mean, in as much as you're willing to talk like about your internal stack. I mean, I was curious A. I mean, you're doing a lot with the Python SDK, and I think that's probably more because that's what people are, you know, outside, you know, especially engineers and geologists, are more familiar with Python; and say that, you know, cuz half the other companies have C#SDKs and they need a actual developer to do that, whereas maybe the Python you can, you know, engineers and geologists can actually tap into that as well. But, I mean, what are you guys using internally and like you know, are you using any like you said, are there any like time series databases that you're using that have been, you know, you know, really helpful? I mean, just cuz I'm sure if some people are trying to like, how do I tackle these kind of solution these these kind of issues. going to follow up part two to that question is can you talk about again at as much length or or not as as you want, but about how y'all do handle like just different it could just be strategic. You don't even have to get into details. But, strategically how you go about, you know, being able to handle hundreds of gigs of data with the real-time component and needing action and models. Like, y'all using serverless where you get spun up and spun down. Kind of piggybacking off of the database piece, but curious on kind of how y'all Mhm. your thoughts on that. So, just to quickly start about Python, Python is a great language. I'm, I'm not a coder by education, so the only language I can code is Python because it's simple and there are so many I mean, ChatGPT can code on Python. And pretty much over the last, maybe, you know, when it was released like 6 months ago, ChatGPT's coding for me. So, uh it's it's it's amazing, it's great, but uh uh Python is pretty much useless for true software development, right? Because when it comes to heavy data or like nice and uh actually what any I think any software company spends 90%of its time is interface and and back end. Uh I would say maybe even interface is more than a back end. So, you need something fast, something reactive, something, you know, uh super easy to use. Mhm. I mean, Python is useless there. Yeah. Python is is great for prototyping. Uh so, as far as uh uh um coding, uh most of our front end uh is done on C++, which is by far the, you know, the fastest and the most advanced uh programming language. Okay. And so, that's for the desktop application? That's for the desktop application, yes. For for web applications, uh of course we use JavaScript. I don't think there are any alternatives there. Back end, we use uh Java. And we have a lot of uh uh other maybe smaller It's probably hard to find a programming language that we do not use right now. Uh we um uh we we use, uh you know, a lot of them, some really old ones, but for, you know, specific specific purposes like maybe some heavy heavy math uh calculations, uh we use one language and, you know, something else we use a different language. Now, uh going into the uh database space, um uh as I said, there are many, you know, databases appear like literally every every day. Uh and uh all databases and yeah, it's We tend to generally look at at what what's going on. Uh specifically uh couple of years ago, we started uh developing uh applications for uh real-time drilling data analysis. And we investigated lots of different databases available. And again, the the the key is not to have like every hour update or every 30-minute update. The we want it to be real-time and analytics needs to be available real-time. And uh we use database called uh ClickHouse. Okay. It's uh fairly new, uh but based on pretty much all performance. The And it was developed, of course, not for oil and gas, right? It was developed for IoT. Uh based on all the performance tests that uh we've seen and we repeated some of them, it's uh by far the fastest one. Yeah, I've heard it's heard it's super fast. It's almost like comparable to say to Snowflake, but you can run it anywhere kind of. Yes. Yes. And uh we tried it on real drilling data, of course, for for that type of database, drilling data is almost nothing. It's, you know, it's uh it's a small data set for them. And uh it's it's amazingly good. Yeah. So, this is really cool. is our uh for now, right? For now, it's our database of choice. I mean, but I'm I'm imagining you're probably architecting that API layer or whatever like to be like somewhat agnostic to the, you know, the database so if you need to rip out one database and pop in in something better, you can kind of Exactly. well. Yeah. Yeah. Yeah. And And again, speaking about deployment outside so outside of, you know, AWS and major cloud providers, very commonly, there is a uh request from operators that they would like everything to be deployed on a specific database. Okay. Because because they like it. I mean, it's it's usually not not a logical reason for that, but uh So, for for for those, yeah, we we're basically ready for those type of those scenarios. This is very important for us. Yeah. I mean, just one thing, just pique my interest. I imagine have you all looked at like DuckDB recently like uh much recently? I mean I I've just heard very good things. Almost like SQLite but for analytics and you know makes for some processing. So just curious if you've investigated it yet or not. I don't think so, but maybe something for us to take a look at. Yeah, and that's interesting cuz you can do through Python or R too is but it's like basically like what SQLite is but it's columnar and you can you can chew through data super fast apparently. So just curious if that's something you guys had evaluated yet but got to have my my chance to nerd out so. Mhm. Tell me cuz we've got we're going to be wrapping up here in a second. I've got one normally we kind of ask before we get into this last part, you know, what's one thing you're excited about as far as the the energy space moving forward or what's some advice you would give. I I'm very curious on how obviously with my my soiree into edge computing stuff. What is, how do you see edge, kind of um,

where, where do you see edge today and how do you see it, kind of, in the future as far as its potential impact on oil field field operations, specifically? Because to me, that's where it has the most potential opportunity, also the hardest thing to probably implement as well. Um I think it's it's a very good question and maybe a further long discussion on edge computer for oil and gas. Um uh Long story short, I think I can understand edge for uh edge devices for uh manufacturing facility or maybe for for for a vessel that you know transports LNG and has no or very poor internet connection. I think for the modern uh and we all understand it's very expensive technology that is hard to deploy and takes a lot of time to deploy. I think for the modern drilling with internet connection speeds that are getting better and better, right? And we are again we're drilling in heavily developed areas. So typically we're drilling in somewhere where internet is is cheap and and good. Um not that edge is not useful, but I think the that component of calculation at at the rig in order to reduce amount of the data that we're sending back to the office. I think it's it may not be that critical and that big. It will be some calculations and maybe data compression on the rig side, but then the data will be sent to the cloud where the calculation is taking place, where we can integrate all possible data or nearby wells or you know, whatever we have. And the final outcome sent back to the rig. So, I think this is where it's all going. Yeah. As far as automation, drilling optimization, you know, just during automation. This is a hybrid Yes. setup almost. Yes, I don't think it will be like those serious serious calculation taking place at the rigs. Yeah. Yeah. Yeah, I mean that's interesting I mean perspective cuz I mean I know there's a lot of people who are making the contention that now there is enough horsepower to run cuz now you can basically run clusters. You can have like a say a cluster of IoT devices that you know, are almost like your own mini cloud. on locations. I know some people are some of people are thinking about it totally differently, but I mean I I tend to lean your way. But it's matter of like that's where you I think it's cuz I know what we've heard say on the completion side, it's like where maybe you don't have a strong internet connection. If you lose connection to the cloud, then basically those people are in the dark for, you know, however long that that connection's out. So, I'm curious if you guys ever run into that. very much, but um Yeah, it's just interesting to hear it, but I mean it sounds like you guys have a pretty strong connection most of the time. And then I think it's best to offload some of that to the cloud and then send it back and you can send it to everyone else internationally or nationally much more easily from the cloud as well, so. Yeah, it's a it's one of those things to me that the benefits of it on paper sound really good. The execution of it is significantly harder, especially cuz it's I mean it's still I would call it fairly new even though it's just on prem again, but it's distributed on prem in this instance, which makes it a lot more interesting. Um But yeah, no, I it's it's a I don't know, right? Like the goal of all of this would be the rig steers itself, right? It's got you dump all the data into the edge device; it's got all your surrounding wells, all the seismic, all that stuff. And it's just running its model and control, you know, integrated directly with the uh, the PLCs and the control systems in the, the rig, and it just does itself. I I think that's uh it's going to take us a long time to get to that spot purely just from the uh risk perspective, right? Like that's the big thing. a lot on here, just so It's it Well, it's it's one of those things though, cuz it's a double-edged sword, right? Like if you're trying to automate, you can't have it solely in the cloud because if you lose that connectivity and the rig it just stops. Right. Right. But then you can't also rely like it's a it's a tricky that intervention. Right. So Well, yeah, you also have to have reliability of that this model is going to work Yeah. all of the time without a human being being in front of it. So, it's uh No, it's a it's an interesting spot. I was just curious, like I said, because there is a lot of the the promise of one of the big promises over the last, you know, 5 years with machine learning, right? Is that, "Hey, it will use machine learning to make our operations better, to automate things, all this stuff." And then, of course, with all new uh technology, they go through that hype cycle of over-promise, under-deliver, and then everyone's really down on it, and now we're kind of coming back up to realistic expectations and My favorite analogy here is uh Tesla autopilot. Think about Tesla, 4,000 data engineers, and unlimited amount of data, and by far data quality much, much better than we have from downhole tools. Yeah. And uh still, if you've ever driven a Tesla autopilot, it's not great. No. Uh it cannot drive by itself. And we will not be able to drive by itself probably next 5 years, 10 years, for sure. Yeah, I just want it when I'm commuting to and from the office, when I'm sitting in traffic, so I can just mess on my phone and not have to worry about running into the person Yeah. of you. Those things are really nice. Well, that and just when it literally gets a mind of its own. I was reading last night there was a story about the US Air Force did the simulations with an AI enabled like drone pilot thing. And but it had a human telling it whether it could kill or not kill, but like it was given its objective was to kill these things. And but what it started to realize was that the human would tell it not to, and it tried to attack the human or the base. That was cuz it, it interpreted that as a threat because it, this was stopping me from doing my job. From completing the mission. Yeah. Yeah, no, that's the the I think the big thing people don't realize about any AI model is it's the the point 1%of outliers that you can't comprehend or conceive of until Years if not deca-like uh one of the Tesla stories I remember, you know, in Tesla like they've done a ton around edge. They have their own in-house developed with I think Nvidia edge device on or edge computer on every single vehicle, right? Like micro you know, Ford and some of these guys aren't doing that. cloud-based edge device. Right. That's the beauty of like that idea, too, right? It's like uh federated learning potentially of, you know, having all these devices learning locally, sending up and accelerating the learning of the mother model or whatever. But uh you know, like there was one it ran into like uh I think it was just like a a white box truck because like and it showed the the Tesla AI computer vision video of it and it didn't recognize that it was a truck because it was just plain the whole thing was white. Right? Like and so like just little [ __ ] like that. Like there's so many of those instances that if you don't train the model, it doesn't know. Yeah, and you take for granted what what your brain is calculating just every instant. Like you just don't realize all the inputs that are coming in that you until you tell it's not there and you're like, "Oh wait, that's weird." You know, like Yeah, until it runs into a another car, right? Um I've got a couple quick We're going to we kind of do at the end uh just we call it like lightning round. Just quick quick answers and or quick questions and answers Um and Bobby will ask you some but I'll ask you some as well. What what kind of software app service have you used recently that you're like excited about that that you've recently found that people may not know about or or even may know about but you're just really like that new tool or whatever. Uh specifically for oil and gas I think it's hard to say It doesn't have to be oil and gas specific. It could be just tech or like I I I mean specifically about I mean specifically about tech I'm I'm super excited about Tesla applications. Uh I hate how it drives but and I won't have one because of that but as far as the software development what they do the true agile approach I I was shocked. I haven't driven Tesla for for a while but recently I took it and I'm I'm just the way the way they reinvented the wheel by by doing software application for a true hardware device. Mhm. That's that's amazing. Yeah. That's a no that's again another underrated piece of tech right? Um what's one of your favorite books either in general or about technology? Uh I like I like to read just books about business business experience. Uh uh just sometimes historical books but mainly about mainly about business and how I wouldn't I probably cannot say a specific one but just people people who build something big and sharing you know what kind of things they went through. However, you know we always need need need to take into account that it's always a survivor survivor bias right? It's it's whoever what what I really like to read I couldn't find that book if maybe someone can recommend I would be so happy to to read it is someone writing a book about his or her failures. You know what what went wrong and why it went wrong. Yeah. So for some reason people don't like to share that experience and this is the most valuable Absolutely. Yeah. I'm trying to think I mean I've read a good bit of business book. I mean it's still kind of of a success story cuz it's interesting hard words, but like The Hard Thing about Hard Things by Ben Horowitz. I think he's got talks about some of the ups and downs at least over the Right. over the time of kind of standing that up, but um but yeah, yeah, to your point like most people aren't going to write the postmortem, right? Yeah, yeah, yeah. That's why I really, you know, respect like Equinor when they did their postmortem with Price Waterhouse cuz it was like, "Here's everything we did wrong on US shale." Like, you know, "Here was the price decks we used and here was all of our, you know, cuz I again people can learn a lot from that." Yes, yes, but it's very painful and yeah, it's you have to have a very specific culture to be able to publish that, yes. Yeah, absolutely. Um Yeah, so I mean what what's your favorite cloud to work with? I mean, probably AWS. I I wouldn't be very very specific here. Um it's it's good. They they they listen to uh they listen to you. They they do improvements. We see those improvements are happening. We have some services we're happy with. Uh we we talked to them. They they accepted it. They, you know, didn't disregard the fact that it doesn't work very well, so they and they modified it and now it works better. So, yeah, I I like I like working with them. Although it's a big company, they they listen to you. Nice. Mhm. What about what's your favorite open source uh tool or service? Maybe other than Blueware, I guess if you guys have used that. Um um Blueware is is not open source, but it's free of charge to use. So, for for those who want to use data compression. Uh I would say generally open source uh I mean, everything that is coded on Python. I bunch of those like you know, GitHub Python repository. Sometimes you just randomly find something that was built for for geology and uh Do you have one that comes to mind that you might uh share, a library or repo? I I like uh so, there is one for uh anti-collision calculations called Well Ankh. Uh well engineering shortened. Uh I think it's just one I I don't know if he's a human person, but like one drilling engineer as a Uh, in free time he develops and uh, he keeps updating it. Um, I I just I I just I just like it very much. Uh, there are uh, other libraries. There is one for us I I actually use it a lot for seismic uh, for seismic data that was developed by Equinor. Um, forgot the name, but I think it's only one seismic library for within Python. Uh, there is one for well log visualization called welly. I also use it uh, on Python. Um, probably those are like the most common ones for geoscientists. those are the kind of things we want to share with our That's exactly what we want to get out in into the world. I mean, it's so crazy to me with Python, right? Cuz there's so many libraries for so many different things, and then you start getting into the oil and gas stuff, and it's like there's one. Yeah. Well, thank god there's one, right? Yeah, yeah. There might not be any, right? Many thanks to those guys. Yeah, yeah. I would say like Python's the second best language for anything, basically. Yeah, yeah, yeah. No, I and that's I mean, that's I love that about the energy community, too. We solve problems, and if we identify that there's not something there, someone at some point will probably come and find find a way to to build it. But um, I'm going to ask: what uh, what is what's something you would, or piece of advice, or something you would tell somebody that's looking at kind of starting their own energy tech company that you know, you've learned over the years, that you would give to them? Um, I know that's a loaded question. Yeah, yeah. Cuz there's a lot of answers to that, but that's not the short answer. Yeah. Yeah, what's what's kind of one big thing that that you would advise, especially around this you know, software side? Uh, I would say uh, the biggest advice is if you think that what you're building is absolutely amazing exciting technology, make sure it's not only you who who thinks so, but uh, you actually uh, challenged your your technology with with users. And uh, yes, it doesn't have to be 100%of approval rate uh, from you know, your future user group, but maybe you just interview, you know 10 10 separate people and make sure that out of 10 at least two love it. Those would be your first innovators. If nobody likes it out of 10, probably it won't fly. Yeah. Now that's a big that's to me that's one of the hardest things of the development in general is you never have like a 100%consensus on pretty much anything from your user base, right? And so it's very hard to like prioritize and figure out or like you might have an 80%consensus, but then there's the other 20%that get mad about it. And it's just like, well, what do you want me to do? Yeah, it's never going to be good enough, but you definitely have to like you said challenge your your thesis. Or I mean again or or you or you can spend too much time working on your baby and then you put it out in the market and everyone tells you your baby's ugly. Yeah, like you said, you just got to kind of shorten those feedback loops and again maybe get it to a 10 people that you trust at least that you will break it to you nicely, but yeah, you got to do that early. So Yeah, yeah, exactly exactly. and pivot the direction. That's very typical problem with any startups. You solve your own personal problem that you would like to solve, but maybe it's only yours problem. Nobody else cares about it, right? So talk to people you know, make sure you know, they have the same problem at least. Yeah, I understand. That's good advice. Yeah, cuz a lot of people go and build and then try and find product market fit. Oh, that's not there. Yeah. And it's okay. I mean generally saying it's okay that people are not I mean most of the it's only what like 5 10%of innovator like true innovators out there. It's okay that your technology will be liked by minority of the people initially. That's totally fine, you know, it will it will be picked up later on. But that minority has has to have to exist, right? Yeah. Yeah. Talking Crossing the Chasm here. Like that's one of my favorite that's that's my answer to that book question is Crossing the Chasm. It's a great one. You got one more, Bobby? Uh Yeah, let's finish. All right, that's kind of easier one. What's what's your favorite video or board game? Um I played maybe too much uh probably boring answer. I played too much Monopoly with my son. Uh Um, so, that's that's we we it's I think it's it's a good game. going to say it's not a boring game. It's very strategic. It sounds simple, but there is a strategy behind it. And I figured it out much later than he figured it out. So. And how old is your son now? Uh, he's 10. 10, okay. That's awesome. Thank you for coming. I appreciate it. Thanks for having me. great having you on. Great conversation. Where can where can people find you and and learn more about Reggie? Uh, definitely website reggie.com, which is easy to remember. It's two co-founders named flipped. So, uh That's funny. We have we have our office like five five That's that's our typical question. Can you guess what's what's the name? But, we have our office here in Houston, you know, 5-minute drive from here and pretty much any place where people drill for oil, we have office or representatives. So, we would be happy to talk to you to hear. Yeah, tell us about the greatness of Star Steer, of course, but please tell us about your pain first. Tell us what, you know, bothers you. We would be happy to solve it. Awesome. Well, thank you so much. Thanks, everybody. Yeah, thank you, guys. We'll wrap it up. Again, go and give us some reviews, subscribe where where you get your podcast. We'll see you all next time.