Episode Transcript
[00:00:00] Speaker A: Foreign.
[00:00:11] Speaker B: Welcome to the Victory Show.
[00:00:14] Speaker A: Hey victors. Welcome to this episode of the Victory Show. If this is the first time you're joining us, I'm Rachel League with Bestseller By Design. Our founder Travis Cody is the best selling author of 16 books and we've had the privilege of helping hundreds of business consultants, founders and entrepreneurs write and publish their own best selling books as well. Through that journey, we've discovered a fascinating pattern. Most businesses really struggle to break past the seven figure revenue mark. On this show, I sit down with some of the world's most successful CEOs, leaders and business owners to uncover the strategies they used to scale way past that mark so you can do the same. So get ready for some deep insights and actionable takeaways that you can implement in your life and business. Starting now. Today's guest is Rajesh Iyengar, the founder and CEO of limcode Labs, a deep tech company using AI and machine learning to help manufacturers detect defects on the assembly line with greater speed, accuracy and confidence. What started as a single conversation with a large manufacturer turned into repeatable scalable solutions that now serve clients around the world. Transforming traditional factories into smart factories through visual intelligence. Rajesh's career spans over a decade of hands on leadership in AI and ML innovation, cybersecurity and industrial IoT. He's also a board advisor, angel investor and startup mentor known for his ability to solve overlooked problems with elegant technical solutions. As an operator, Rajasth is deeply passionate about turning pain points into platform businesses. He spent the last eight years building linkode into a go to solution for quality management at scale. And we're excited to hear how he's done it. From productization to partnerships and from ideas to impact. Rajesh, welcome to the show.
[00:01:52] Speaker B: Thank you so much Rachel. Thanks everyone.
[00:01:55] Speaker A: So take us back. What first got you excited about AI and cybersecurity and how did that eventually lead to founding lyncode?
[00:02:02] Speaker B: Yeah, absolutely. So you know, after my graduation I've been in the space of data centers, so I worked with various data center distribution companies or either solution companies around the data centers. And with that experience I got into a lot of data center designing or servers or even cybersecurity and so on. So that piqued my interest in doing more deeper into those. And that's when I got in, you know, touch with about what is AI, how Google is using elasticsearch, when you go on Google, what are they using in the backend, how is it done and things like that. So that curiosity got me more deeper involved in the AI space, you know that that's how the journey started. And then when I did my couple of previous startups in Silicon Valley, more so on the AI space. Lynncode was my third startup with AI and a few of my manufacturing customers approached me asking for solutions using AI because they knew that I come from AI background and they said like, can you help us? That's how sudden.
[00:03:02] Speaker A: Amazing, amazing. Tell us a little bit about how you knew that this thing that you were passionate about could be a scalable business. And what were those first steps you took to transform it from those conversations with manufacturers into something that was official?
[00:03:19] Speaker B: Yeah, absolutely. So when we started in way back in 2017 with AI, I was not sure I was going to go only in manufacturing. So I tried out with retail space, I tried out with healthcare and then finally I chose manufacturing. And the reason for that is healthcare. Once you create a data set with one particular patient profile, be it your X rays or dental or any of those healthcare industry, it becomes very competitive or it becomes very monotonous because once you create a data set with certain people then it's easy for you to go and you know, do things with AI with similar data set. Basically you don't need to collect data set for every hospital, every clinic, it's going to be common basically across across the world. Whereas manufacturing is very unique. Every manufacturing industry is different, every manufacturing lines are different, every factory floor is different and even the automations for each and every floor is completely different. So in such scenario things get very hard and very tough and creating a AI model for us, each and every customer on a scale is impossible. So you need to come out with very different ways to actually solve this and go around and solve this basically. So that is something which we wanted to do and that's the interest that we had to help the manufacturers. And there were a lot of mission vision industry which is basically little bit earlier era of AI. So they were basically focusing more towards identifying defects only with like you know, checking the presence and absence, checking if the feature is there or not. Like for example if the logo is present or not or if there is a screw or if there is a fastener is there or not on a component. So this is something which they focused on. And also the next one was like OCR optical character recognition or even reading text or even a barcode or a QR code. So those are the technology they were focusing with, whereas these were not solving the customers 100% problems. The main devil was in the surface defects. So when I say surface Defect, it's basically like scratches, dents, some kind of a burr or things like that. So what happens with this is these are very complicated to identify because you need it's only the defects are only visible from certain angle with certain lighting conditions. And if you turn on the light in the evening, when the factory is in full production, daytime it works fine, but nighttime, when you turn on the lights, the shadow is falling on the components. You're not seeing it clearly. So this other common challenges everybody goes through, that's why, you know, we felt that this is much more complicated to solve than any other industry. So that's why we dived deep into this.
[00:06:07] Speaker A: And if not using blink code, would
[00:06:09] Speaker B: it all be done manually, partially manual, partially. It would have been technology.
In the industry they always do the secondary inspection, which is a human. So with the machine vision technology, they could not 100% rely on the results of it. So even after the inspection is done with technology, they had to do the secondary inspection for a complicated components, basically. And they could not do it for every production components, they used to only do batch testing. So out of one, the whole batch for that particular day or particular shift, they checked few components out of it. Like if they produced some hundred components on that batch, on the batch, they would test only about 20 components. So that's the time they had. Otherwise they could not go to all the hundred and test everything basically. So this is something industry follow and that's what's going on.
[00:06:57] Speaker A: And given you mentioned earlier that there was a real challenge with scale different from other industries, where there's a common floor map with manufacturing in particular, what were the first signs that you felt like you had something that was quite repeatable and scalable with your solution? And you know, what steps did you take to turn that into a very scalable product?
[00:07:18] Speaker B: Yeah. So one thing in the automotive industry is even though the component is almost identical, or in fact in some cases it's the same components, even then that way it is processed completely changes between two companies. Company A is selling a component to a oem, let's assume, and company B is also selling the same thing, they both will have a different process, different automation, different type of procedures itself. So because of which, even if you have identified a way to solve one specific manufacturer, and if you want to use the same thing to other manufacturers, it's near impossible because since the process changes, they may not do the same way it has done in the previous factory. So when you're going to deploy, you may have to change everything according to the new company's way of things, basically. So you can't really fit everybody into the same solution basically. So you have to have a change. And that change is the biggest concern in the industry. And that's why a lot of AI companies or machine vision companies don't want to do it for every customer, like, you know, in that manual way, because they cannot scale otherwise. So that is the area that we solved it by a very unique way, basically using our platform. So we combine multiple AI tools which can actually, let's say if I roll out a model to one specific customer and we go into the other customer, we start from zero, we won't have any knowledge about the previous customer. There are various things, right, like, you know, copyrights. We don't want to use one customer's data to, you know, use it in other customers data and so on. So when we go in, we start from scratch. But what we do is we basically collect a few samples from the customers and then we look at how this can be solved. Like we do a consulting on how the camera and lighting has to be taken care and how do we, you know, ensure it can work in an open environment. Because today in manufacturing lines it's also very critical that you don't eat up more space for the quality inspection because they already are space constrained there. And if you demand that, you know, I need enclosed space, I need a lighting very close to it, I cannot do in open environment. This kind of creates a hurdle for them into going into production with that. So we have to be very friendly in the manufacturing floor where we actually look at, okay, we are going to be working on open environment or retrofit into our existing line, which is basically, let's say if there's a robot or a automation process which is producing the component, we can retrofit the cameras and lighting to work with the existing space itself. So that way they don't need even another area to do the inspection. The inspection can happen or coexist when the production, production is happening. So that's the first thing that we did. The second thing is we created a model on the go based on the AI tools that we have. We are able to get to a very high accuracy model within a couple of weeks, which is unheard of in this industry.
And we are able to achieve that because we have experience of almost 8 years now working with different AI models and different factory environments. So we fixed it. It not just with the software. Hardware is also very key here. Like the way you set up your cameras, the Lighting, the connectivity, the existing legacy systems of the manufacturer, and then finally the software piece as well. So it's like software is just AI, is just one piece of the puzzle. So this, everything combined together becomes a important thing for a customer.
[00:10:39] Speaker A: Yeah, it sounds like the technology was really the advantage in being able to serve different customers scale both on the hardware and the software side. And tell us a little bit the team that you've built around you to be able to continue to evolve the technology and meet the needs of your ever growing customer portfolio.
[00:10:59] Speaker B: Yeah, sure. So we have a fantastic AI data science team which is basically like we have. We understand that, you know, in the AI space, innovation only can solve many complicated challenges. Like we have done a lot of innovative stuff. Lot of R and D has gone in both hardware, optics and even the software as well. So first thing is, if you go into any customer, they don't care what technology you use. All they care about is can you solve my problem in a real quick time. The second thing is how much of savings your technology is going to help me out with or your solution is going to help me out with. Right. So with that kind of approach, we need to be innovating quite a lot to make sure our customers don't pay a hefty price for, you know, acquiring a technology. So first thing we should be focusing on is lowering the hardware cost, lowering the kind of camera cost that they are investing in. And the most important thing is it has to be a longevity as well. Like when they're going to invest in a camera or a lighting or a hardware, it cannot be just for a year. And then next year you go back and say that, oh, you need an upgrade. Right. The whole capital cost goes for a toss. So you have to make sure that you consider the next five years. And even if they are going to build another new line or a new product, they should be able to reuse most of the components without wasting any money there. So that's something it's very important to do in the R and D stage itself. So we spend a lot of our money and time in R and D. So we have a fantastic data scientist people we have in house mechanical engineers, we have in house electronic engineers who all combine together to combine come up with a total solution for our customers. So we have a CPO who's managing our product and he's taking care of the AI part of it. And then we have a VP of engineering who is handling the front end and back end and user experience. So we have a great UX team as well, making sure the product is not just giving results. It's also easy to operate as well. So that's the important piece that we look up to too. So we have a great team on that front and then we have a amazing DevOps team. DevOps is a very important piece of team which is required for you know, companies like us because it involves a lot of images, lot of, you know, lot of companies do a production and they sometimes like to use our cloud because they have six months to one year they need to keep the data and if they maintain their own environment it becomes very expensive for them. So we try to reduce the cost there as well. So we run those on the cloud. So we need to have a fantastic DevOps team to make sure that we don't increase our cloud costs. Basically we have a great team.
[00:13:37] Speaker A: It sounds like you've built on a very large team. How have you resourced the business from a human capital perspective and of course a technology and R and D perspective over time have you raised any external funding or has it largely been from profits from pre existing customers?
[00:13:55] Speaker B: Initially when we started it was bootstrapped since I had a couple of exits we could sustain with my own money, my own investments initially the first two years and then later we raised the venture capital first. So that's an interesting story. Do you want me to share my first ever how I raised my money? So I was based in the Bay Area, I was operating out of one man army, operating out of a co working space and they used to conduct a lot of events and I used to go on and you know, just listen to those events. I was not charged any fee or anything. Used to witness and meet a lot of different people and then one such such event was a startup poker event and I never played poker ever before and it was the first time for me to play poker. I learned, you know they, they said, you know it's only for exclusively for the founders plus C you know investors, you should enroll yourself and any other. They used to conduct this every month in the Silicon Valley, once monthly, once. So I said definitely, I want to be part of this. I love this energy of all the founders. I love the idea of meeting fellow founders every month once. It would be fantastic experience.
So the next event I went there and when we are introducing ourselves like the format was before you start playing poker you introduce yourself and you can ask like you can what you are, what you have and ask and what you want to give. So there's a like quick 2 minute roundup from each and every founder and investor who comes to that event. So when I was saying that what I do and what I'm looking for, there was a guy who was from vc. So he heard me and he was in my table, in the poker table. So he said, hey, why don't you come and meet me tomorrow? He gave me his card and the very next day I went and met him. And within a week from that meeting he said, we are in for the check basically. So they were our first investor. So that was a fantastic experience. And I became a strong advocate of this startup poker because it's basically helps you connect you to with the fellow founders, learn from their experience, learn from their hardships, how they came up with a startup idea, what did they do, how did they come up with all the, you know, how did they cope up with all the issues typically a startup goes through? And then you, you have a nice game as well, which is, which is fantastic. Which teaches you how you can run your startup as well. And then you can meet any VCs as well. So it's a fantastic one. So I became a great advocate of it and I approached the organizers and said I want to be a big part of it. So they made me the chapter director. So I was one of the chapter director for Silicon Valley and now I moved to Michigan. So we are going to start our first ever poker maybe next for the first, very first time in Michigan.
[00:16:34] Speaker A: Amazing. I agree, it's. You can't build in isolation. It's absolutely a community and a team sport regardless of what you're doing. Of course there are solopreneurs out there, but even just idea sharing, I think building in public is such a great way to move faster than you could on your own. And it sounds like you had a really positive outcome just from showing up to something that maybe you'd never done before. But I love that you're leading the chapter now in Michigan and you've been CEO of Landfield now for eight years. You had run prior companies, successful exits. How has your leadership style evolved across your entrepreneurship journey?
[00:17:13] Speaker B: Yeah, so, you know, even Steve Jobs would have always said that, you know, I keep learning from everybody, I keep learning from my team. That's why he used to work closely with his team and not just sitting around in his cabin. So, you know, that's something I really follow as well. So I've done lots of mistakes right from the beginning and even now. So I learn from those mistakes and see where we can improve. Work with the team very Closely and listen to all the team members, even if it's an intern.
We even give a lot of weightage to the interns because we learn a lot of new things because they come with a completely new perspective, fresh perspective. They're viewing from outside first, and then they're coming into the system. So they bring in a fresh perspective as well. So we do learn from them. Yeah. So it's a great learning journey so far.
[00:18:01] Speaker A: What's next for lyncode?
[00:18:02] Speaker B: So far, I would say we have not yet reached the scale level. We are probably growing. We want to take it to the next big scale. Basically, we see that we can really grow into a very large company. So there's a big market. Potential is very big. There are tons of opportunities.
It's just that we need to, you know, focus more and get going with the scale.
[00:18:24] Speaker A: What does victory look like to you today?
[00:18:26] Speaker B: Victory? You're asking, like, how victory will be for me or today as a victory? So far, what we have achieved is the victory you're talking about or the victory that we are going to achieve in the future.
[00:18:37] Speaker A: How would you define victory?
[00:18:39] Speaker B: Okay, so victory means achieving some goal, you know, which is more tough, and you have achieved that goal. That is victory for us. So far, we have achieved smaller goals, smaller victories. We have seen, like, for example, we got the deal from one of the top automotive manufacturers and initially getting inside them was a hard one for us. But when we overcome our competitors and won the deal in the US with this automotive major automotive. So it was a big victory for us because some of our team thought we will never be able to achieve it.
But some team members did believe that it is possible and because of their faith, we could achieve it.
[00:19:20] Speaker A: Overcoming immovable obstacles. I love that. If you could give your younger self one piece of advice, what would it be?
[00:19:26] Speaker B: That's tough. I think lots of. But if I have to say it is, you know, it might sound a little bad, but it is what it is. So, you know, trust people less and trust the process more and fire bad people sooner.
[00:19:41] Speaker A: Yeah, it sounds like your. Your intuition, your gut feeling is. Is a key piece of that. Well, Raj, thank you so much for joining us on the victory show. It was an absolute pleasure.
[00:19:51] Speaker B: Thank you so much. Ra.
