Masters of Technology Happy Hour
Conversations with masters of technology, those who produce it or those who use it.
Masters of Technology Happy Hour
S1 Ep 5: From Rockets to Design Revolution: Sohrab's Engineering Journey
Ever wondered why design software hasn't kept pace with our AI revolution? Sohrab Haghighat, aerospace engineer turned entrepreneur, is tackling this exact problem with his startup Hestus, bringing genuine intelligence to CAD systems that have remained fundamentally unchanged for decades.
The frustration will sound familiar to anyone who's designed hardware: you have a brilliant concept, but getting it from your mind into reality means endless hours wrestling with test equipment design, manufacturing constraints, and assembly relationships. As Roopinder explains, "The computer-aided part wasn't in Computer-Aided Design." Hestus aims to change that by creating AI that thinks like an experienced engineer, anticipating your needs and automating the tedious aspects of design work.
What makes Hestas' approach unique is their focus on practical, manufacturable solutions rather than the flashy but impractical "blobby shapes" often produced by generative design. Zorab points out that many current AI design tools create structures that look impressive but prove impossible or prohibitively expensive to manufacture. His team is building AI that understands manufacturing methods, tolerancing, and assembly requirements—essentially putting a manufacturing engineer in the box with you.
The potential impact is enormous. Hardware development cycles that typically stretch over years could be compressed to months, allowing engineers to focus on creativity and innovation rather than documentation and constraint management. For companies trying to bring products to market faster and more efficiently, this represents a genuine competitive advantage.
Beyond his professional life, Zorab plays soccer with his seven-year-old son and does ballroom dancing with his wife, a chemist who became his co-founder in a previous rocket company venture.
Ready to see how AI should be used to transform your design workflow? Listen now to explore the future of engineering tools and discover how smart sketching is just the beginning of a complete revolution in using AI in CAD.
Hello everyone and welcome to the Masters of Technology Happy Hour, where once a week, I have a drink with someone I meet in the course of business, but someone I'd like to get to know better as a person. So today I'm joined with Sohrab Haghighat. Zorab is an aerospace engineer by education and experience, but what's most interesting to me is that lately he has started a company called Hestus and he's putting AI into CAD, which is, I think, in my opinion, long overdue. Right, I think CAD has existed for oh 40 years now, just trying to live up to its name of computer-aided engineering. The computer-aided, I've come to realize, wasn't there. We're going to do that. That's the plan. Okay, that's the plan. Good, and, by the way, cheers, I don't have a real drink today, but because this is a happy hour, I have water to drink, let's toast.
Roopinder:Maybe one day we can do this in person with real drinks. So tell me about your journey now. How's it going, how's your week and how's the company? You're making great progress in the company, so tell me how it is Today's going fine?
Sohrab:It's been quite an interesting week that we had, with a lot of focus on stuff that we have built so far. How can we continue with the momentum that we have had and keep delivering on the promise that we have to our users that we are going to make their lives easier our users that you're going to make their lives easier? One of the reasons Hestus is a startup is because of the frustration that Kevin and I have felt over the past many years as engineers ourselves and different roles. Kevin Chu is my co-founder.
Sohrab:He is a longtime friend and my co-founder at Hestus.
Roopinder:And I think how did it work out? You were actually working with him or for him, initially, correct, so, Kevin was the person who hired me at Cruise.
Sohrab:So, if I'm not mistaken, Kevin was employee number five at Cruise Automation and by the time that I joined the company I was employee number 10. And for two and a half years that I worked at Cruise this is before acquisition by GM and also I continued working after acquisition by GM. I worked very closely with Kevin. Post my time at Cruise I left to start my previous company, SpaceRyde, which was a very different journey and very different business of building space launch technologies. But, as I was saying, both Kevin and I come from hardware background and have had our fair share of dealing with the frustrations of designing, testing and deploying hardware. And the main goal that Hestus has is to make it easier for hardware developers, in particular mechanical engineers, systems engineers, test engineers that have to deal with designing, testing and deploying hardware.
Roopinder:Got it. So what your startup is doing now is trying to help engineers make it easier for them to do design work. Correct, okay, so let me back up for our readers' sake, because I know this from talking to you before. But your company is actually making it easier for solid modeling. You can actually sketch easier. There's more involvement with AI into inferring what you're trying to make with sketches. So it puts in things like symmetries, one-axis or two-axis symmetry, you call it mirroring. It helps you with that.
Roopinder:Ordinarily, sketching does not do that much in ferring so you've made sketching smarter, correct.
Sohrab:So it is correct that where we have started is to make sketching smarter and to be able to count engineers that part of the design, which is basically the first step or the basic step that engineers go through. But the ultimate journey, or the ultimate vision that we had in mind and we are still working toward that is how to make a product that streamlines the process of mechanical design and hardware design. Helped us understand this pain point more than anything. Was the need for designing test hardware, peripheral hardware. So every time when you design a hardware, one thing that you always have to do is to test it, and often times to test the hardware you need to go and build yet another hardware, and that task by itself almost always takes equal amount of time. The testing aspect of the design takes equal amount of time than the design itself, and oftentimes you are in such a rush to get the design of test equipment done, to get to testing, to figure out what are the issues that you need for them to go back to the drawing board to fix your initial design. So that puts a lot of pressure on systems engineers, test engineers. It always shows up as a disruptive in terms of scheduling for mechanical engineers, and we thought that, okay, if we automate this whole process, significant burden and load is lifted from the fate of these three groups of people mechanical engineers, test engineers and systems engineers so they can focus on the creative aspect of the design.
Sohrab:Designing a frame, designing a jig, designing a mount by no means is the funnest part that one wants to do. It's a necessity that you have to do so. If I take it away from you so you don't have to deal with that, you would be stoked that you have more time to deal with the creative aspect of the design. Oh, this robot that I'm designing has to function such and such. How should I shape the design so we can do that and be the most optimum form factor in doing so, as opposed to oh, I need to test the robot. Now I need to go and design a test rig. Sure, it is a cool thing, but that's not.
Roopinder:You don't go to the party because of I can't wait for this to happen, because all the time when I'm trying to design something, I get mired in that drudge work that you're describing. I get, I have a creative idea and I want to run with it, and I get mired in having to first of all, create it in CAD, Second of all make a prototype and third, test the prototype. I just don't. That's just to me, it's just. I just want the damn thing to be out there. Yeah, Exactly. So what you're doing is, you know, every discussion of AI is usually in the background of this discussion, there's always this fear that AI is going to take my job. But what you're saying is AI is going to help you do your job right, oh, 100%.
Sohrab:What I'm trying to accomplish here is not that oh hey, you can do this with your engineers. No, I want to make sure that if, for example, a program that typically runs three years a hardware program and maybe a good two years of that, is a testing and back and forth and iteration, I want to cut that two years into months with the same team, so they can get through more development and more products coming to market. So team is not stressed out, they are not going nuts because they have to juggle so many things that that is actually not helping them. They have to do all that in their head, thinking about manufacturability, thinking about testing requirements, thinking about assembly requirement. How should I design this product to be assemblable, so one part is not blocking, for example, the whole access that I need to secure this other component?
Sohrab:All of these things today is done manually and is a source of stress and friction at hardware company. All we want to accomplish is to get rid of that, to literally free engineers from doing these mundane tasks. That is not contributing to the performance of the product, to the quality of the product. Help them focus on what really matters, the conceptual design, the creative aspect of the design, and that is the vision. But the question that also matters here is that how do you achieve the vision? And that's why we started by Sketch Helper, using AI to help engineers with sketching, and I claim this is how we are different than literally anyone else who is thinking of using AI for hard writing.
Roopinder:I got to ask you this thing and I think I did ask you this, but the name Hestus comes from. Is it a derivative of the name Hestus, the great god? Of fire? Craftsman. Okay, got it. Thank you.
Sohrab:So that's how he came up with the name. The credit is actually for Kevin. That's how he came up with the name. I was saying how much this is different than any other AI, CAD or AI hardware developer is that I've seen everyone is focusing on oh, you can use, I don't know, boxes or enclosures for Arduino or this thing or that thing, and someone else is doing oh, you can use this text to do this niche application. The way they are approaching the problem is how a human designer is actually thought about designing a product.
Sohrab:When you go to school to study mechanical engineering, to study aerospace engineering, don't start day one machine element design. That's not how you start. Don't start day one machine element design. That's not how you start. You start by studying statics, dynamics, how load is transferred, how that load puts something into motion. Then you start looking into strength of material, how things bend into material. Now, strength of material, too advanced version. But after that you go into machine element design you don't jump into. Okay, this is how we design gear.
Sohrab:Tomorrow I'll teach you how to design a box, cause that is endless type of stuff that you can design. You cannot instill creativity like that. You don't teach them the. You basically don't give them a fish, you teach them how to fish. And that's exactly what we are doing. We are starting by sketching, then we are going to progress into the next step. What do you do after that? Cheating, modeling. What do you do after that? Handling assemblies? Okay, once you do these things, then you can say okay, once you have an understanding of how things are assembled together, how things are put together, what are the relationships, the types of joints we're making that matters here then you can put all of these different layers of AIs or you can look at it as sketch agent, 3D agent, assembly agent together and put that to work beside something from ground up that is not limited only to examples that it has seen, that can do more than just a box for Arduino or I don't know, just 3D design, mesh generated gear that if you don't like it, you cannot even change it.
Roopinder:Yeah, that makes sense. So it's a lot more than sketching. You'll eventually be able to work with things that I'm used to making. I'll give you an example and I'll make a request, if I may. This is just me asking for things that I want, so of course, I'm just another voice out there. All the time when I'm riding my bike. I all the time when I'm making riding my bike, I ride my bike a lot and I have often thought of my bicycle as like a perfect geometry for that purpose, that diamond frame bicycle made out of tubes. And all the time I think is this that design? And I've asked a few CAD companies too that are convinced that they can design better shapes than humans through generative design, to come up with a better design my diamond frame tubular construction bicycle and they've all had a go of it. Or they've decided to ignore me and they say, no, okay, we don't want that problem, that's not what we're trying to do, or they make up all sorts of excuses not to do that problem. And the answer is obvious to my mind what is the optimum solution, but I can't get that.
Roopinder:place where a cat can actually helpfully optimize a design by using shapes that CAD familiar with? Like, I don't want to start with molecules and make a blobby shape that's not going to be a bicycle frame. I want to use tubes. Tubes, engineers know, are perfect in torsion right, and they're good in tension and compression right. . And use tubes. I don't want compression. making blobs of blobby shape ridiculous looking frames, start . ridiculous-looking frames, to or use tubes.. the shapes that I'm familiar with that I want to use? If I'm a civil with,, I want to use I-beams right. I love I-beams right, they're perfect for some cases. If I want . They're house, I want to use 2x4s because that's what's available, right? I don't want it to start making 3D printable shapes.
Sohrab:So that's actually taking me back to one of the core principles that we have is that, as I mentioned, tad today has no understanding of certain fundamentals. One of them is manufacturability. What does it take to make something manufacturable, given the method of manufacturing that you have in mind?
Roopinder:Right.
Sohrab:Now let me give you an example For my previous company. We were designing space technology, specifically rockets, and for part of the rocket engine we wanted to shape these aluminum domes, this aluminum dome by the way, I should mention SpaceRyde. yeah
Sohrab:When I was working at SpaceRyde, one of the things that we wanted to build was this rocket engine that the company custom designed, and the dome of the engine was supposed to made out of a specific type of aluminum alloy. And our idea was that, instead of machining that from a solid block, we can actually get a plate of aluminum and spin form that into the dome that we want. I'm not sure if you're familiar with the process of spin forming, but that's one of the manufacturing methods. One thing that we didn't know is that alloy that we had selected had too much hardness that we could easily shape and form it using spin form. This is something that my engineer didn't know using spin form. This is something that my engineer didn't know. This is something that we did not know until we did the design, went through the fabrication, quite a few iterations with a manufacturer who had been expert in spin forming and eventually failing, coming back to the point that okay, there's a specific alloy is too hard to deform in this method. That is something that I want to create that level of awareness that, okay, if you are using CNC, if you are using injection molding, if you are doing spin forming, if you are doing 3D printing.
Sohrab:Each one of these technologies and methodologies that we have for manufacturing have certain limitations and have certain requirements. You, the designer today, have to be aware of literally every requirement and every principle for the manufacturing method that you choose, to make sure, by the time that you finish your designer's manufacture or the way it has come. At bigger companies there are entire departments designed for manufacturing that, asked the fact work on okay, how are we going to actually make this? Is it doable? Whereas this is where we see AI being useful that by the time that you finish the design already has solutions that make it or has basically a requirement that makes it manufactured well, given the method that you have? So back to your point why everyone takes your request of bicycle and gives you either nothing or gives you a shape that okay, how am I going to build this?
Sohrab:Because that manufacturability is not top of the mark, right, and part of the reason we think mesh generation, for example, is also not something that is suitable for mechanical design is because if something looks good to you, to your eyes, it does not matter how it is generated. If I have taken a blank slate and I have changed every little pixel until it looks like a rabbit, or I have actually gone through the process of thinking what is a rabbit? A rabbit is an animal. It has two long ears, it has this cute nose, it has whiskers, it has teeth that are like this and it draws it. To me my eyes, doesn't matter if it was generatively designed or if it was drawn ground up, basically understanding of what the rabbit is.
Sohrab:But when it comes to mechanical design and hardware development, it actually doesn't matter. If you come up with something mesh generated, there is no guarantee for that to be manufactured and there's no way I can.
Roopinder:Actually, it's impossible to manufacture something like that, so a lot of times.
Sohrab:So either it takes a very excruciating detailed terms to get something and I don't know how big of a model it takes for all manufacturing requirements to be enabled and verified or you have to be able to modify that which for a mesh generated solution. How are you going to modify that? The whole beauty of parametric CAD is that you can go and change this parametric, enlarge it, maybe shrink it, change how much you extruded something, or the depth that you cut something, or the angle that you revolved something. None of that is available to you. That's why, instead of going down the path on text to can we actually looked at okay, let's teach AI to cap properly Sketching 3D assembly this way.
Sohrab:Once we enable requirements that, okay, these are part of the requirements for manufacturing 3D printing way, CNC way, injection molding way, sheet metal way. Then every step, you can try to satisfy those and you have means to satisfy those because your AI knows how to do that. So, one of the things that you mentioned people have been giving you blobs, and these blobs are not by any means new things. Today they're being labeled as generative solutions. I haven't seen any solution for the past 20 or so years, back when I was doing my PhD. This is what we used to call topology optimization we used to call topology optimization.
Sohrab:Yeah, and for many years people have been talking about, oh, airplane wings that are being designed like that, in funky shapes, or basically wing ribs. Ribs are these structures along this stand of the wing. That basically gives it.
Roopinder:Not a day goes by, I don't see a new wing structure or airplane structure. Linkedin yeah, it's amazing.
Sohrab:It's all beautiful and flashy looking, but the question is how?
Roopinder:are you going to manufacture that? How are you going to make that the manufacturability is not considered right? Yes, and that's why people have been giving you weird looking bikes.
Sohrab:But the goal is, once you have this product built, be able to give you a bite that you can put together from bits and pieces that are available to you or easily manufacturable. So the part can be made.
Roopinder:So it's like putting a manufacturing engineer in the box. This is your manufacturing assistant. You're going to get this Like. Even it could be a young engineer. I don't know a thing about manufacturing and you're saying this could be a function of an AI program.
Sohrab:That's under a person what you want to know.
Roopinder:I gotta say, though, I got thrown out of a machine shop early on in my career because I went into that. I was exactly that engineer I just described. I went into a machine shop with my drawing, thinking whoa, I'm a great engineer, just got out of school, I know everything about everything. And I gave the drawing to the guy in the shop and looked at it, scratched his skin and said not exactly these words, but where's this effect? Get the hell out of here, come back in when you have a real drawing. And I couldn't figure out what he's talking about. But when I looked at it I realized. And another experienced engineer told me you can't have four-place death suits on these holes.
Sohrab:You just made a million-dollar part, you dummy, that's actually one of the things that a lot of time. For example, tolerances that are too tight, because engineer is either afraid that part is not going to fit together properly or they don't have an overall context of how things are going to fit together in a whole assembly. So they just don't want to do stack of tolerancing for a massive assembly, so they opt for the tightest that we can get here and there. Of course you can manufacture things like that, but cost is just going to go to the room. Build a product successfully.
Roopinder:Build it. Yeah, you can't think assembly. How much AI could benefit assemblies, Not just with tolerance, stack up analysis, but also with like in SolidWorks. You have to spend a lot of time deciding meets and just to make sure the parts come together properly, and all the time I think, oh, shouldn't that know how it's going to come together? Shouldn't you know here you have a round hole and here you're going to need a screw? Shouldn't it know that this screw fits there? Or you're going to need a weld here, or something? That should all be obvious 100% 100%.
Sohrab:Basically, you import two parts into your design. It should be able to recommend what type of joint or borrow from SOLIDWORKS terminology mate, slide, and if it is, for example, I don't know evolving, what is the range that it can move? So, instead of everything being specified by the designer, these are areas that you want to provide help so you can reach through documentation of your design in CAD. All the designer is doing in reality is documenting what is in their head in the CAD software. So if you think of grammar, for example, like how they're saying oh, we help you improve your writing, make sure punctuation is correct, grammar is correct, it is very concise. That's what we want to help you with your CAD work. So the concept that is in the head comes onto the computer screen fraction of the time that you do it and comes exactly as you had in your head wide satisfying manufacturability requirement, wide satisfying assembly requirements, testing requirements, all of these things.
Roopinder:I've got to say this is pure genius, but you also have to know I attribute genius to people who agree with me, but I really wish you the best of luck, because this is music to my ears. I wish it happened sooner in my career so I could make use of it, but I'm glad to know it's happening now. What are your challenges, though? Because, even though this is a grand idea and I totally get it I think the industry needs it. How, it's not going to be easy to do, because the companies that I know are doing it well, that we all know that are doing AI are big companies, right, big companies, and only one of them has even come close to addressing AI for engineers, and that's you've heard of who Jensen Wong, who talks about physical AI, that AI is based on physics, that it understands objects can't go through each other, that they have to interact. Are you making use of that sort of AI?
Sohrab:So definitely every tool that we can get our hands on we would be using Okay. But one of the things that is making our technology defensible is, more than anything, AI that we are also developing from ground. So there is a great deal of work that has happened in the world of LLM and their create or other AI models that people have been using. But one thing that we truly believe is that it's what we are building is for a certain and very specific application. Rather than taking giant model and share their tunic to have, okay, we can actually build a small model from ground up that is desired to perfectly solve the kind of problems that we want to solve. So we get stellar performance out of it. And when it comes to inference and running, that is actually not becoming a bottom or a cost issue for our customers.
Roopinder:Okay, all right, got it. I promise we wouldn't talk about work, and then we ended up talking about work, but I actually liked this steps. I hope it's. It was fun for you. But before I move on, tell me something else. What else in your life would our viewers be, our readers or listeners be, would find interesting about you? What do you like to do in your spare time when you're not working, for example, what's an hobby?
Sohrab:One of the most fun things that I do. I play soccer with my son. I grew up playing soccer, fascinating that, without us really pushing him in that direction. We tried to expose him to as many sports as possible, so he became obsessed with soccer. So the fact that we have our interest line means that every weekend, when I get an opportunity to play with him, this is one of the most fun things that I do with my son, and another thing that a lot of people actually find it surprising about me.
Roopinder:I don't have this plan. by
Sohrab:Another hobby that I really enjoy doing is ballroom dancing.
Roopinder:Oh, wow.
Sohrab:Yeah, so that's something that my wife and I picked up many years ago, so when I was younger and we started dating. My wife has always enjoyed dancing, so she signed up for a dance class at the University of Waterloo. Back then we were doing our graduate studies and I mostly accompanied her because I wanted to accompany her, I just wanted to spend time with her. I had no passion for ball dancing, but over many years this has turned into something that I just enjoy very deeply, and the moment we are on the dance floor I feel like time comes to a complete stop. Everything stops and I can just enjoy myself and being in the company of my wife and just moving around on the dance floor freely, not having to think about anything else. As far as I know, the warmth can come to an end at that moment and I would be at peace with it, and that's. These are the two things that I do mostly.
Roopinder:So that's like your yoga right, so you can relax and drift off into another world and take a little break even yeah, nice. You talked about University of Waterloo. That's Waterloo in Ontario.
Sohrab:Correct, yeah. So I went to school in Waterloo and University of Toronto as well. That's where I did my PhD in aerospace engineering and I was primarily focused on this concept of multidisciplinary design optimization. The idea is that when you are designing a complicated system like an airplane, a multidisciplinary system like an airplane, you cannot just go and say that, okay, aerodynamics people. You go to your room, design the outer shape of the plane. It's structural dude. You come up with the thickness of members, hey, control systems. You figure out how the thing is not going to go unstable and come crashing down.
Roopinder:You've got to all work together For a very long period of time.
Sohrab:That's actually how things have been done, but at some point people realized that, okay, if we blend together the design of aerodynamics and structures and it's called aerostructural optimization we can come up to a better performing aircraft in terms of mass goes down and aerodynamic efficiency goes up.
Roopinder:You're talking about how the wings actually became fuel tanks.
Sohrab:Not that. Mostly you have a thickness distribution that gives you, for example, the stiffer wings. If you look at it alone, or if you look at, for example, aerodynamic efficiency alone, which would pull you into different directions.
Sohrab:But when you look at it from the same perspective of okay, what is my optimum solution and what should it look like Ultimately, I want to maximize endurance or minimize, for example, fuel costs Then you would end up at something that looks like none of the structural only optimum or aerodynamic only optimum. But it is giving you a better performance in terms of what you were asking for. Endurance, fuel consumption range that it can cover. So that's basically the new aircraft shapes that you see and how it has changed from what it was back in the 50s or 60s. That is the outcome of this type of optimization that is being done. When I was doing my PhD, what I was thinking about was that, okay, we primarily size structures in many ways to withstand gusts and turbulence. That is maybe 10% of 90 % of the flight duration, flight duration, but for you're carrying this extra mass that you don't need.
Sohrab:Now what if you use active control technology to relax the structure, to alleviate this extra load when you are encountering with it? When you are hit by a gust or turbulence, the control system comes into effect. When you are hit by a gust or turbulence, control system comes into effect. And basically what I was working on is not just combining aero structural optimization together but do what used to be called aero servo elastic optimization. So servo. Referring to controls technologies, so how controls can come into play, referring to controls technologies, so how controls can come into play. And basically my focus was working on these high endurance style surveillance planes that could fly for days, or even the idea was to push it to a point that you can achieve. So that was basically what I was doing at University of Toronto.
Roopinder:Got it Now. You met your wife there. She was also a grad student.
Sohrab:I met my wife at University of Waterloo. She was a graduate student.
Roopinder:Yes, you're not studying engineering, is she?
Sohrab:No, she's a chemist by training. She studied analytical chemistry and for most of her career, when she was working as a chemist, she worked at pharma companies or medical device companies and then studied MBA at Haas School of Business at UC Berkeley. And then later, when I decided to start SpaceRyde, my launch company, I could not find a better co-founder than my wife. So my wife and I started the rocket company together and one of the funniest thing was that one time when we had a parent-teacher conference, a teacher asked us that, oh, we asked everyone about their parents, basically what they do for a living and whatnot and the parents said something that we couldn't quite make sense of it. And my wife asked did he say that my parents make rockets? And she said, yeah, he was telling you the truth. It's strange. That may sound, that's actually what you do for a living.
Roopinder:You're actually a rocket scientist. How do you like that?
Sohrab:Yeah.
Roopinder:He sounds like he's proud of you, which is good. So what's happening? I'll let you go because you're not making anybody talking to me, but this has been a lot of fun. I hope we can have another one, because I think we have a lot to talk about If you'll only listen. But are you going? What are you gonna do? Is there a ballroom dance or a game today?
Sohrab:No, no ballroom dance or soccer game, but my day is mostly talking to users to figure out what is working for them and what else they want, to see how else we can help them, or do as much work on the development and see how else I can help the team. We are a very small company. We are a team of three and, as such, everyone is handling so many different things. As you said, it's a very tall order trying to build AI for hardware development from ground up and changing your own model, and I'm really excited by the challenge that we are going to. Basically, we have set out for ourselves and tried the tasks that we are trying to basically work hard, but what it means is that you got to shift from one thing to another thing very quickly to make sure that we are addressing all different parts of the business.
Roopinder:That's great. Okay, this has been good, very good. Hey, I wish you the best of luck. It's very hard. It's very the event diagram of hardcore engineers and those with entrepreneurial skills that can actually do something with it. It's a very small intersection, so if you're in it, I encourage you to do all you can. It's a big path ahead.
Roopinder:If you ever need me for recommendations, I'd be delighted to give you one in this effort because, it takes a long time, but it takes somebody who is really an engineer to understand what engineers really need, and there's few people, that few engineers are actually willing to advocate for their profession to have this happen. Most of them are like, happy enough with their tools or they don't really question it.
Sohrab:So good for you. Thank you and really appreciate for taking the time to chat and rooting for us, and all we hope is that it would make hardware development less and less painful and users experience more joyful, so people can go through more desires. Creators get to work on different products throughout their life and their career, so there are more things to point at and say that hey, mama, that's what I did, be proud of me, so there are more of those products out there.
Roopinder:Go make mom proud. Good, all right, all right. Bye for now.