Dhruv Sharma: Hey there, listeners. You're watching The Offline Network. We're streaming live. This is stream number 142. And today we're speaking with Pradeep Chaudhry of a company called LeanWatts. Pradeep, welcome to The Offline Network.

Pradeep Chaudhry (LeanWatts): Thanks, Drew. Thanks for the introduction.

Dhruv Sharma: Great. Where are you joining us from, Pradeep?

Pradeep Chaudhry (LeanWatts): I'm from Hyderabad.

Dhruv Sharma: And, I believe you you are seated in your R&D facility-cum-factory right now.

Pradeep Chaudhry (LeanWatts): Yeah. That's exactly correct. Yeah. It's a it's a newly established, facility. We've just started it a couple of months back.

Dhruv Sharma: What are you guys building over there?

Pradeep Chaudhry (LeanWatts): So, we work on building power electronics. So, just to take it a little deeper, so we work on EV chargers, hybrid inverters, and, power modules required for the energy industry.

Dhruv Sharma: Fantastic. So you're a power electronics company that, at this point in time, has started by making EV chargers.

Pradeep Chaudhry (LeanWatts): That's correct.

Dhruv Sharma: Do you sell—I would imagine you sell the chargers to OEMs?

Pradeep Chaudhry (LeanWatts): Exactly correct. So, we have a couple of base where we sell our products. Currently, OEMs are one of our biggest customer segments. And on, their side, we sell to, eventually, we'll go to OEMs, but, through battery pack manufacturers, through some kind of network where, but it'll end up in a vehicle. So that's what, that is a product we make. Yeah.

Dhruv Sharma: Alright. We'd love to learn more about, you know, chargers and charging technology. So maybe let's just start by drawing a distinction between, onboard chargers and portable chargers.

Pradeep Chaudhry (LeanWatts): Yeah. Sure. So, I'll just add one more thing to that. They're called public chargers. So I'll explain the difference between all three. So, think of it like this. You have a two-wheeler and a car, an EV vehicle. So, if you're using a charger, which is outside the vehicle, it's generally called a portable charger. So imagine this, you can keep it in your bag, keep it inside the vehicle, and carry it across. And the other kind of charger is called onboard charger. So this is, what is more prevalent in four-wheelers. There should be a default onboard charger in your car. So this is, what takes the AC port from the grid and converts and, charges your battery pack. So, yeah, the name itself, says it. Right? It's an onboard charger. It's on the vehicle. And, there is another kind of charger called the public charger. So this is, where, essentially, you can't, stuff all the capability of, how much your battery can accept in the vehicle. The reason being the weight will increase and, the range will come down. Especially everyone is optimizing for the range right now. So you can always opt for, keeping it outside, which is what the form factor is. So, this charges at, like, 10 x to 20 x the speed of what an OBC can do.

Dhruv Sharma: So those are, like, more like fast charging, kind of devices. Interesting. And so you were talking about the, the onboard char charger. Does that have power conversion units, like, that that that amounted on the charger itself?

Pradeep Chaudhry (LeanWatts): Yeah. That's correct. So, I just need to do one more, distinction. So there is something called as AC charger, what people generally use it. So, yeah, that is why where the confusion might stem from. So in the industry, that is actually called as electric vehicle supply equipment. It's not in, charger as such. So it's just a switch and safety device. So that also will look like a charger, but it's not a charger. And, the OBC, which is what the charger, the terminology is, this is actually what converts and the power conversion technology lies inside that. So it it takes the power from the grid, carefully does a lot of things, and, converts from one form of electricity to another form of electricity.

Dhruv Sharma: You're getting AC from the grid, but, obviously, there's a battery that that's gonna work on DC, so that's why you need the conversion take place. Interesting. And can you, also spend some time explaining to us what makes it particularly hard for batteries to operate in India, the climatic conditions, and also, you know, electrical conditions, fluctuation, and so on.

Pradeep Chaudhry (LeanWatts): Yeah. Yeah. So in general, this this, is true for all the factors, two-wheelers, three-wheelers, and four-wheelers. So, we are a country where, that we have temperature extremes going from anywhere between minus 10, minus five to 50 degrees in summers. So that is one, huge decision factor when we are designing, any kind of product, for outdoor use, particularly. So, essentially, we'd have to take in the account of humidity, moisture, rain, and, yeah. These are the essential things. And on the other side of the, thing, like, we, unfortunately, even after, growing such a huge power infrastructure in the country, There are few, there is huge imbalance, and fluctuations in the power supply part. As everything else, this also has to make sure it will take its own time to make sure. A mature power grid will have, less surges, less, fluctuations. So, essentially, you're trying to, operate, all your equipment when your electricity is fluctuating all the time.

Dhruv Sharma: Yeah. No. It happens. I mean, many, many years ago, you for all your devices, you would just have installed stabilizers with them. You have no option but to. Interesting. And, I'm assuming that in the peak of summer, if you're using, you know, fast charger, then you're putting a lot of thermal strain on the battery. Is that something that's not generally recommended?

Pradeep Chaudhry (LeanWatts): See, Yeah. I'm, not from the four wheel industry currently, but, yeah, I can weigh on this question. So, it's it's it's, engineering design. It's related to engineering design. So, essentially, there is no such thing as thermal strain or any kind of strain on the battery if it is designed correctly. So, essentially, what happens is, each of this equipment has its own operating conditions. If it is going beyond that operating conditions, the vehicle will take a call and not do that task. So it could be overcharging. It could be over temperature and, or, the in fact, the products what we make will have, at least 10 to 15 different types of safety cutoffs, hardware level, software level, at a system level. So, in the car, it's even higher. So, yeah, to answer your question simply, it's advisable if the OEM okays it because they'll be doing extreme testing for this for months and months and years. In some cases, the platform continues for ten to fifteen years, so they'll have enough knowledge to allow something and not allow something. It's, yeah, in India, the particular problem is more relevant because we don't trust our OEMs for whatever reasons. Some cases, it's not reliable. Some cases, we expect more from them in general. We want them to reduce cost or whatever it is. Yeah. It's, it's it's okay to fast charge it somewhat. It's not a problem anymore.

Dhruv Sharma: Will you expand into four-wheeler chargers at some point?

Pradeep Chaudhry (LeanWatts): Yeah. Yeah. So currently, the products, what we make is for two and three-wheelers chargers. Company, we are really, really revamping the entire organization. We are adding a lot of R&D, activities inside the company. We are making a new lab. So we'll be simultaneously releasing the products for, four-wheeler fast chargers and hybrid inverters in the coming, six months to one year.

Dhruv Sharma: So you view yourself more as an R&D company?

Pradeep Chaudhry (LeanWatts): Yeah. That's that's exactly correct. So our focus, although, we are currently dealing with, the building, the company, managing the customers, and R&D at the same time, but we do want to focus more towards R&D, because of all the points what we just discussed. There is a lot of development that is required to develop, the right products for the right audience, in this case, our country.

Dhruv Sharma: Alright. Pradeep, now that you've given us enough of a primer on chargers, tell us also how, you know, you and your founding team, came came about to build this company and not some other company.

Pradeep Chaudhry (LeanWatts): Yeah. So, so, we are three cofounders. So I've been working in the EV industry for the last seven, eight years, also including this, around ten years. So, I have another partner like, called Abhilash. So he used to run a a home electronics company before this. So he understands electronics. And, during my stint at my previous companies, what I've observed is, yeah, we, I was including myself. I was focusing more on the systems and, bigger picture, like, because that is what we have to start with. Right? You have to start with the car first, right, in this case or an example. So then I realized, like, you know, with this, urgency for speed, urgency for cost, we rely on, improving subsystems components. In this case in my case, I've seen the charges generally. Most of it was important at that time, when we started this company. So that was a clicking point. And then once we, entered into discussions, we understood that, everything, any industry which is growing right now needs a power component. Yeah. So that's when it clicked and we started the company. And then the the third cofounder called Sujit, so he's a school friend of my, cofounder. That's where, he also got interested in, and now he's leading the sales and customer, development, part of it. And, my first cohort at me and Abhilash, we met in, our college, BITS Pilani.

Dhruv Sharma: Wow. All of you know each other from one campus to the other. That's, always great. Pradeep, can you highlight, you know, technology breakthroughs along the way that made what you're selling now possible? And, also, why are OEMs buying from you guys and, you know, not from someone else who's lining up to to supply to them as well?

Pradeep Chaudhry (LeanWatts): Yeah. The yeah. I'll catch up on the first point, first. So, when we started, like, see, obviously, it takes us some time to time to took us some time to get a grip on what we were doing. So, essentially, to figure out what technology is gonna work, in our case, it's a power electronics board, the hardware and the design part. So, just to, not a couple of technical points, like, what topologies will work for this industry because that will drive the design, that will drive the component costing, the end costing, the serviceability, the lifetime of the product. So I think that is, when we mature, when we understood, okay, we are not designing the product for building the company. We are building the company to serve the customer. Right? So we understood, like, you know, at certain cost, they'll buy it. Certain cost, they don't buy. At if you're, supply or in in our case supply, and essentially, you should serve your customer on time when they need it, not when you can make it. Right? So this is when we got matured and, the business understanding is what, is a tipping point for us. And, on the competition side, I always have this theory, like, you know, it's, you can buy from anyone. That's true. But, customers will buy from us particularly because we focus on each and every aspect of the, product design, supply, and the service part. So, again, we learned it from experience. So we are able to piece together everything in a a symphony, to be able to supply it to the customer on time. And when there is an issue that happens, we are able to support them on time. So, yeah, essentially, we've compressed the 12-month development process to a six-month process because of our own technology in house and India-built stuff. And on the service also, we took the same, approach, and, we have built a a portal where they can, raise a complaint within couple of days. It'll be resolved by default. Otherwise, we are gonna replace their product. So that is a promise we made. So, it's it's a combination of, really, individual links in each, every department, which, makes our customers choose us.

Dhruv Sharma: Right. Are you talking about after-sales service?

Pradeep Chaudhry (LeanWatts): Sorry?

Dhruv Sharma: You're talking about after-sales service?

Pradeep Chaudhry (LeanWatts): Yeah. Yeah. After sales is, really, really important in our industry.

Dhruv Sharma: Yeah. Talk to us about the product portfolio, Pradeep. What products do you have right now? Talk to us about their specs as well. Yeah.

Pradeep Chaudhry (LeanWatts): Sure. So, we have currently, products in production for, two-wheelers and three-wheelers. So, essentially, onboard and onboard chargers for, two-wheelers and three-wheelers. Imagine in a company like Ather or Ultraviolettete or, any of the two-wheeler manufacturers, when you buy that vehicle, the charger, what you get along with it, could be onboard or could be off-board, is the product what we build for two-wheelers. And the similar product for three-wheelers, in terms of ratings, we designed from 500 watts to 6.6 kW, range. And, we do have multiple variants available in in that inside that as well. So, essentially, each vehicle works on a different voltage range. So the battery pack might be, requiring a 48-volt system to 60-volt system, 72 and 96, which are the common, voltages people are using in India. The companies are using in India. So, essentially, we have, yeah, this combination and combination of all these roughly 20–30 SKUs, we offer in the market right now. Yeah. And, the next generation of, LeanWatts, we are adding another business vertical inside LeanWatts. So that vertical will be taking care of, high power systems. So imagine your hybrid inverter for your home, imagine hybrid inverter for your battery storage systems and power converter for public chargers. So we, yeah, essentially we supply the core heart, which actually pumps the electricity, converts and pumps the electricity into the systems.

Dhruv Sharma: Interesting. Pradeep, talk to us about how battery chemistry itself is changing and how you foresee, you know, charging infrastructure in a sense having to keep up over the next three years, five years, and maybe even globally, not just restricted to India.

Pradeep Chaudhry (LeanWatts): Yeah. So, globally, what happened in the last fifteen years was, Tesla has taken the, NMC cells from, Norway, to a place where, everyone has used it at some point in their, company's lifetime. And it has become so huge that, battery production has tripled, even quadrupled in the last five years at itself. So, the king is, these two technologies, the NMC and the LFP technologies. So NMC is generally used for high power applications. And, LFP is used for high energy applications, where you require power for a lot long amount of time and, slowly you consume it. So that is the major, technology which has created all these possibilities right now in across all the industries we are planning to work in. So LFP is the main, the technology. And finally, there are other things which are just at a nascent stage. See, when I say nascent stage, it has already crossed the R&D stage and just started entering the industry. So the my apologies is coming. And, yeah. So, commercially, if you take, these are the three things which are relevant for the industries right now. Apart from this, there are a lot of varieties which, will be required for niche applications. For example, if you want to, take very high lifetime applications, you can focus on, titanium based charges. And, if you want to go in space application, again, delivery will become a very key part. So there'll be some other very niche technology which goes inside that. So, yeah, essentially but it's a combination of, these three which will be relevant commercially, especially for a scale, like, Indian market.

Dhruv Sharma: Yeah. I would imagine your R&D Lab tries to lay its hand on all kinds of batteries from all around the world just to understand, what's coming next. Yeah. Pradeep, what how does it feel to be building, you know, an engineering first deep tech company right now? Talk to us about the round you've recently raised.

Pradeep Chaudhry (LeanWatts): It's a roller coaster. So most of the times, we enjoy, like, taking decisions, which will impact the next product design, next product release, and, all, basically that. And, on the lower side, like, with, it's it's a constant struggle to, you know, keep working, for get the right capital at the right amount of time, get the right talent and, make sure that you stay in the company. And, also on the customer side, make sure you engage them at the right trend and, not fight for every day in the cost on the technology. So we'll have to find this kind of things. So, overall, it's it's a pleasant journey. It's a lot of learnings for me and for the team. And, yeah, it's it's a journey, and, it's not a a one day job.

Dhruv Sharma: Yeah. Well said. I think just final question for you for today, Pradeep, as, you know, the founder of a power electronics company that's making EV chargers, what tips can you give EV owners in prolonging their battery lives?

Pradeep Chaudhry (LeanWatts): Just use your car like how you use your ICE vehicle. Don't bother too much about this. Because, once you know that, that fitting your lifestyle, just enjoy the vehicle. That's the best thing you can do. Any part, is will fail. Any system will fail. Don't worry too much about it. Start enjoying the car, like, what it has to offer.

Dhruv Sharma: Are there tips, like, they give you for for your cell phone batteries? Like, don't charge it all the way up to a 100%, pocket in the shade and stuff like that.

Pradeep Chaudhry (LeanWatts): These are very sophisticated systems, which are, as I told you, there'll be multiple levels of protection, and, these are already taken care of.

Dhruv Sharma: Sure. Yeah. Alright, Pradeep. Thank you so much for joining us this afternoon. All the very best to you and your team.

Pradeep Chaudhry (LeanWatts): Thanks, Drew. Nice talking to you.

Dhruv Sharma: Likewise. Great. Alright, listeners. We're now gonna be welcoming Sandeep of DSW. Sandeep, welcome to The Offline Network. It's a pleasure to have you.

Sandeep (DSW): Thank you very much, Drew. It's pleasure of mine.

Dhruv Sharma: Thank you. Thank you for joining us. Please tell us what you do at DSW.

Sandeep (DSW): Wonderful. So, you know, when we started thinking about what we need to build, and it was not just because we wanted to build another AI company, but this entire thing transpired through the ground level reality, what we faced at that point of time somewhere late 2021. And that was, you know, where we were seeing and experiencing that while enterprises are going to build models, agentic was not then in the industry. Models and, you know, and adopting AI. A point is going to come where build is okay given and it should be done. But how do we operate this? Because anything which gets built, it has to operate in a very sustainable manner into the production. It has to get into the enterprise architecture. And this learning we got when we are doing a real-time experimentation for one of the insurance company where we all agreed that, you know, while the pilot was ticked as a success, but nobody was able to take it in the production, not because of technology, not because of the limitation of knowledge, but because we are not sure that what will break when it has to perform continuously. Because in pilot, what happens, you have to be right once. In production, you have to be right every millisecond. And that's where the thought came that, you know, there is a time eventually after four, five years where enterprises, if they have to leverage AI as a capability, not just build AI For their business purposes or business value chain use cases or business processes, they will need to have AI in their core architecture. And hence, this journey progressed through stages. It was never like, you know, in first year, we are able to achieve that. It progressed through evolved through tool to platform and to the operating system where we are today. So I'd just like to just, you know, give an analogy here that today we don't think about electricity daily when we are working on particular thing, but it's a core silent nervous system of our entire thing, industries, offices, everything. If it goes away, everything goes away. The time is going to come probably, you know, next 2–3 years where AI is going to become silent nervous system of the enterprises where people will not work on AI daily, but they will leverage their processes on top of AI. And that's the layer, lower layer, kernel-level layer, what we are building, where we are thinking from architecture level. And that's why we call it as an operating system. So we are building Unify AI OS that price the operating system from India. I think we take pride in that because this is something a new category creation we are trying to do from India, for India, and for the globe.

Dhruv Sharma: Mhmm. Thank you for that introduction, Sandeep. I think it sets the tone for a lot of things that we're gonna go deep into. You know, just as you were saying, the rate of change with AI has just been so phenomenal. Three years ago, the models were hallucinating. We were talking to them literally one prompt at a time, and here we are now with in within an enterprise of humans and autonomous agents working alongside each other. You know, ever since you've started the company, how have you seen the enterprise attention sort of shift, and what is it focused on today?

Sandeep (DSW): This is really very important question, Dhruv. And if you'll speak to enterprises, you'll be surprised that the way we use AI today is not the way enterprises intend to use the AI in their organizations. Because if a tire manufacturing company leverages AI, it's not going to manufacture car anyways. It's still going to manufacture tire. The way they are looking at AI is to bring out the intelligence from the data what they have. If I had to go a little bit back, three decades back till last, you know, few ten years back, Whenever any software we have used, it has processed the data and given you the information. With the AI, entire game is changing. When you process AI and data together, not only information, intelligence is arrived at. And which is something completely different ballgame. Those enterprises are also not aware of the intelligence which is coming out of that. And that is why we say that a time is coming and enterprises are completely aware of this. Not only data, but intelligence is a new asset of enterprises. And that is where enterprises are looking at how they can operate AI not just as a tool, not just as a feature, but how they can become an AI-native organization, which means not just adding AI as a feature to their existing process, but how they can build their business processes at every stage they are leveraging AI. That's complete shift. The way of adoption, the way of building is happening. And in that, every beautiful tools so all technologies are beautiful. Personally, I feel there is no inferior and superior technology. It's about the right fit. All tools are beautiful, but the customer wants control and flexibility. They really want to choose the right fit for their business, not choose and then struggle to arrive at the ROI of that tool. And this kind of shift is an architectural shift. And this shift will need us to look at AI from system lens. And if you need to look anything from system lens, you need to have something which will operate at system level. And if you need to have something to operate at system level, there needs to be a horizontal layer, which is an operating layer for those system of AI. We are not replacing Linux, Unix, or Windows. Those are the operating systems which are governing and computing, governing and operating compute. This is a layer which will help enterprises to build, govern, operate, manage AI. So platforms can be integrated. External tools can be integrated. They themselves can be building building their use cases or workloads on top of this. And the best part to what towards what we are working is whatever they build on this, the source code of all the artifacts remains in their custody because this is a long game for them, 15-year game. They can't just build and get logged in into someone. So I think the shift what you talked about is more of architectural shift, which is enterprises are looking at.

Dhruv Sharma: Yeah. No. That's super interesting. There was, you probably saw this post from Satya Nadella where he spoke about something he called reverse information paradox. He said, you know, you're actually if you're if as an enterprise, you're working with the Frontier Lab, you're paying them not once but twice. You're obviously explicitly paying them for the intelligence that you're renting, but you're also letting them have your data if you're just being super naive about it. So and then he spoke about governed context. You guys use this term called runtime governance as code. I'd love you to do you know, again, just go a little deeper. Explain what that means.

Sandeep (DSW): Okay. So if if, you know, if you have seen this entire journey, and we all have seen this, four to five years back, each one of us, including enterprises and individuals, we got to experience the capability of AI with different tools or maybe LLMs, which are there. Then we started enterprises started thinking that how they can bring these tools in a very structured and standardized way in their organization, that's where the journey progressed to the framework. And when they started looking at the frameworks, they thought that can there be something which can help them right from data integration to building the use cases? And that was the journey we have seen and are seeing is a platform. But then a thought came and when they were in this journey that all this is fine, but at an enterprise level, how can we govern the AI? Which means not the guardrails. Yeah. Governance means suppose if there are 10 agents today are working towards a statement of business purpose, which I would call a collection of use cases, Each agent would have a persona, and they are supposed to consume a specific data. So while building stage, even if you forcefully give it different data, it should not execute. So that is a governance at a part as a design. Not once it is built and then you apply the policy score and governance on top of. So this is a governance of entire AI cycle. Guardrails are something what you can configure on top of that based on your sector and based on your industry. So when we say that, you know, governance is becoming very much important, we say that it should be unbypassable. It cannot be bypassed. So there is a human in the loop in that. And that is where you will see that while AI will be gaining more and more momentum for adoption, at every stage, a human will be in the loop. And that is very important part. And coming back to what you mentioned, you know, you quoted, Satya Nadella that, you know, how this, intelligence is going out. That's why I said, along with data intelligence is becoming an asset. So now every not only organization, but at country level, they are looking at sovereign models and sovereign layer, which doesn't mean proprietary. It means that whatever sovereignty we have, our data and our mesh should able to deal those lens, bring it down, make it more relevant for our industry's frontier models. So there is a lot of work happening in this field. It's so if you'll see that, you know, I would not be surprised. And the way I can see is in next 12–18 months, we would be having a beautiful, powerful, sovereign, you know, frontier models for every sector and industry in our country as well. And when this happens, we'll need a sovereign layer also. Yeah. And that is the layer which is important. So we are building that ecosystem, not just in isolation. Yeah.

Dhruv Sharma: So, Nip, as you're seeing, some organizations, some enterprises are very aggressively adopt AI. What are they doing to keep their AI costs in check? In other words, how are they maximizing token efficiency and minimizing token wastage?

Sandeep (DSW): There are various ways enterprises are at the big AI, and they are learning by themselves also. And this is not about the size of the enterprise. This is about the stage where they're in the adoption of AI game. So if you'll see earlier, everybody tried a lot of pilots using the tokens and hyperscalers where they felt that, you know, irrespective of whether in pilot or production, they had to keep paying the bills, APIs, tokens, and everything. Slowly, now they are seeing that how these things can come to their custody, and how they can leverage which they should not reinvent the wheel like transformers and LLMs. How they can bring it down. And one of the thing that everybody is trying to optimize themselves is understanding what they're building and what should be the sizing for that and what is required for that. So nobody's trying to play the blind game now without understanding what is the infra getting used behind. So that is where if you'll see the enterprise, especially in banking and insurance, if you'll see, they are mostly starting with statement of business purpose. Probably for insurance, you know, one of the statement of business purpose, for example, could be, we would like to increase the persistency of our policyholder by 10% in the next 12-months. This is a statement of business purpose, which is clearly attached to the ROI. So people struggle to derive ROI by AI deployment. No. ROI will come from your business statement. Now once you break this business statement, you will see the processes. These processes becomes your AI use cases. So enterprises are optimizing their AI adoption game by not starting by which model or which AI or which tool. But what is my statement of business purpose? If I enhance persistency by 10%, I am saving ₹350 crore of my insurance company. That is clearly matched. And then they see that how we can achieve this statement of business purpose to the target of 10 enhancement through AI and through the domain knowledge. So it's not a single party game. It's a collaborative game where domain and AI play such an important part. It's not just AI. It's a domain and AI. And that's where they are seeing that things are changing the shape and form, and they are coming out with different processes in that. So that is how they are optimizing.

Dhruv Sharma: I'm also hearing that they're in a sense, along the way, they're calibrating their AI budget to the outcomes that they're witnessing and the outcomes they're seeking Yeah. Absolutely. While spending more. Very interesting. Talk to us more about Unify AI OS, Sandeep.

Sandeep (DSW): Okay. So it has also gone through the journey as I mentioned. So but today, we have, you know, distilled it down to what we call it as a kernel layer like Linux, you know, where there is a kernel, which is a governance layer actually. And in that layer, what we are building is and has been built as a governance of the entire AI cycle, whether it is, you know, agentic or whether it is an ML cycle. On top of that, we are also having an AI fabric layer where we are building an ecosystem around AI, where we say that customers can choose their own tools. They can choose their own external use cases. They can build on top of this, or they can procure from external world. It's something like this. If you deploy Linux today, you can have various softwares. It doesn't, you know, limit you from deploying the softwares. Earlier, it was, but the ecosystem grew that today everything is available on Linux. So every development first happens on Linux. But two decades back, there was a struggle. There was no ecosystem. The same thing I'm seeing with the AI. That ecosystem built is very important. It's not about the players building technology in isolation. So Unify AI is an effort to build that kind of layer where we'll be playing a lot of collaborative game, where we'll be collaborating with multiple, you know, technology companies, multi multiple application companies, and we'll be also building in skills and enablement for the customer. So we are trying to bring that power back in the hands of customer. We are not going to dictate that if you deploy Unify AI OS, this is the only way you should build your AI strategies. Like Linux, again, I go back to that. The moment you deploy Linux, Linux doesn't dictate you. You get the complete flexibility and that entire rich ecosystem to as per your business requirement to deploy on Linux. It doesn't dictate you as a closed source or, you know, as a proprietary in terms of that, you can only do this. So I think so this is way this is the this is how AI is also going to grow because AI is not a point-in-time technology. That's why I always say, AI is not something which is going to process your data and give information. It's going to process your data and something different is going to come out. Intelligence. So it's a different game. So it has to be ecosystem game. So Unify is an effort of building a horizontal sector-agnostic layer wherein enterprises will be able to build their AI-native architecture and become AI-native. And when I say AI-native, today people are adding just features to existing system. Nothing wrong in that. Yeah. But nativeness means which is already there. So intelligence will be there flowing as an electricity. And you build your processes on top of that leveraging intelligence along with the data, not just data. That is a time which is going to come in next 2–3 years.

Dhruv Sharma: Yeah. In fact, we'd love, we'd love to learn more about that as well because with with the kind of companies you're talking about, these, you know, scaled up like, I mean, you can't take a scaled up production system offline just because you wanted you know, overnight, you want to turn it AI-native. It's going to be one Herculean mammoth effort that'll be ongoing for months, if not years. What are the steps for these really large companies to transform into AI-native avatars of themselves?

Sandeep (DSW): Again, I would say this is very intelligent question, and this should be looked through, you know, in a very systematic manner. So when AI is getting adopted, it's not replace something. It's not that an existing database system, which has to be replaced with new database. It's not existing middleware to be replaced with new. So it's not a replacement of the technology. That's the beautiful part of this. It's an addition of AI into their enterprise architecture. So things remain same, but how it gets added, like, for every process, they will leverage AI. So it's not that something you are replacing and putting AI into, but it will run parallelly. That's why it's a complete design of architecture. And trust me, as we move forward, the importance is of architecture. And today also, if you'll see why most of the enterprises are not in production full scale with AI. It's not a technology limitation. It's not the tools limitation. For last six to seven years, all beautiful AI tools are there. But still enterprises are not in production because the challenge is architecture. How to bring that in? And the organization which will work around that architecture will be the organizations who will be able to build nativeness of the AI capability. So nativeness, you had to build the architecture. You had to see that how your processes will flow through AI or how AI will be integrated in your existing processes and enhance it and maybe new processes you can design on top of that. So there is no direct one on one replacement of the technologies. It's a part of architecture. And this is a wonderful time, Drew, I'm telling you. Wonderful times. The people in fact, I feel the people who have seriously worked on enterprise architecture for last two to three decades and who today feel that because of AI, they are not they may not be able to scale at the age of forties or 40 or fifties. Those people have much more potential today because they will able to bring that architectural knowledge how AI should be embedded. So I think beautiful time for new generation and those everyone who has that kind of knowledge. It's a beautiful time.

Dhruv Sharma: Yeah. I can imagine. In fact, thinking of that, Sandeep, are there can you think of examples of trade offs that architects are having to make, you know, short term versus long term? I don't know. Short term, maybe you see a competitor get ahead of you because they're not being as protective of their enterprise context. Long term, you know it's gonna come to hurt them because by then, everyone else will have caught up, but, you know, they would have given up something that they'd accumulated over 20–30 years. So, I mean, I'm just making an example up. Maybe you have better examples of trade offs that architects are having to make right now.

Sandeep (DSW): So one thing I would say here is that, you know, a time is going to come where if you talk about enterprise enterprise level you are talking. Right? This or at the architect level.

Dhruv Sharma: I was actually at the architectural level, but

Sandeep (DSW): Architecture of enterprise. Right? Enterprise architecture level.

Dhruv Sharma: Yes.

Sandeep (DSW): Okay. So this this this is this is also very significant one because here I'm just going to comment I already know that every enterprise has access to the same powerful tools. Yeah. Same powerful technologies. So technology is not going to be the differentiator. The differentiator will be how they are deploying AI capabilities to bring out their own intelligence. Moving forward, every organization intelligence is going to be different, and that is going to be the differentiator. And now that will be dependent of multiple factors. What kind of data they have? White what kind of, you know, business legacy they have? What kind of, scale they have, what kind of customers and geographies they have. So more and more today, you use LLM. Right? So LLM has universal knowledge. What do you do? You ask question to LLM query and it gives you the answer, but it is not very contextual till the time you make it very contextual through applying RAG.

Dhruv Sharma: Yeah.

Sandeep (DSW): Same here for enterprises, they can use all the tech AI technologies to bring general knowledge or, you know, high level knowledge or intelligence. But if they had to enhance their own organization, it has to be more contextual. And that is where the trade off would be, not just tools, but they have to see that how they are going to leverage tool at each stage of the process, not just at the process at the right hand side. They have to see AI getting deployed and leveraged at data engineering, data science level, ML engineering, agentic, workflows, orchestration, and that entire process of factory line or production line or business value chain use case. So at architectural level, the trade off would be some legacy systems, the way processes are happening. That will change. But that change can be eventually coming up. It's not drastic changes. They can still keep running that and augment AI to that. But, yes, lot of things will change. Play field is going to change. The way processes have been built will change. This change will definitely happen, and these changes are constrained. So new constraints will come up. One example I will give, and I heard this example in one of the, you know, very, beautiful conversation that earlier when, you know and you might have also heard this. I think this was by Sangeet Paul. He mentioned this, and I really like that. It resonated to me. In earlier days, if you'll see that, you know, the anyone who knows typing, typewriters, they were paid very well. Okay. Maybe three decades back or more than that, that was a very prominent job. Typist. And even typist used to think that their job is important because they can type so many words in one second. They will they used to relate their expertise with the speed. But eventually, they realized that the value and expertise was that they are not making mistakes because that was the constraint. If you do one single mistake, you have to tear off the entire paper. There was no backspace/delete button. The moment word processing came up, that constraint was removed. That constraint was removed. But still, there were a lot of people who were, you know, who ramped up and scaled to that. So AI will remove lot of constraints of earlier processes.

Dhruv Sharma: Yeah. And

Sandeep (DSW): that will look like, you know, there is a lot of things which are going away. Yeah.

Dhruv Sharma: Yeah. I mean, one of the ways I like to think about this is, like, every profession got its own workbench, a tool that changed everything forever. Like, I don't know when engineers got, CAD or, you know, people who work in simulation, etcetera, got MATLAB. It's like just one piece of software that the before and after your relationship with time changed forever. But all of that happens slowly over a long period of time. It went profession by profession. But AI, like, it's everything all at once. Like, everyone's getting their version of something that's gonna change their relationship with time. So maybe it's this has been a very nice conversation, but maybe as a closing question, you've recently announced a $5 million raise that brings your total financing up now to about, I think, a little more than $8 million. How are you gonna put that money to use? What are the priorities for the next six to 12-months? What is the team really excited about at this point in time?

Sandeep (DSW): No. It's always wonderful, you know, when when you receive funding, that is more than you getting funding. It's the it's it's it's a reinforcement of the trust of people in what you are building. So it actually motivates in terms of the direction what we are taking, and lot of people are now trusting on this, including customers. There are three ways we are looking at, you know, deploying this fund, and we never raised fund in very, you know, hurried way if you see our records where where we could have easily raised in US, but we were not because we are building something very fundamental, which needed us to focus for almost 3–4 years building that. And while building, we acquired customers also. So there are three ways we are deploying this fund. Definitely accelerating what we are building. So strengthening engineering team because this is a very long term game and, you know, this is very, I would say, very critical game because this is going to become a nervous system of enterprises. It has to be right anyways. We can't go wrong. So we had to be we had to ensure that we are not messing up with this. So we'll be strengthening our engineering, our product line. Second, definitely, we are looking at, you know, GCCs and data centers where this layer would help them to build their own AI need to infrastructure. So if you deploy this, they have captive customers. They can start leveraging without movement of petabytes of data. Because eventually, AI may AI cost start coming down, but your data movement cost will be so high that you'll not able to leverage and process the data for intelligence. Second, we are looking at, you know, working with ISVs very closely, especially in regulated industry, where they can accelerate their journey by embedding this layer for AI capabilities. So they don't have to build AI capabilities into that. And third is enterprises. So banking, insurance, and financial BBFSI customers, we are focusing while this is a sector-agnostic. These are the three GTM where we are deploying the fund, but prominent fund is going in building the US region. So we have already incorporated there, and the US entity is a subsidiary of the India entity at this point of time. We take pride to build from India again, and we are building strong team over there. We are in a, we are getting into banking and insurance over there, and that would be our effort of deploying this fund and, you know, investing over there to see that how we can build a new region. And that is US region, what we are building.

Dhruv Sharma: Fantastic. Sandeep, thank you so much for your time, and all the best to you for this super energizing, and and exciting journey ahead. Thank you again for coming on The Offline Network.

Sandeep (DSW): Thank you. Thank you very much, Dhruv. Thank you very much.

Dhruv Sharma: Thank you. Alright, listeners. That was us for today. Again, thank you so much for joining us. We'll see you next on Friday. Cheers.