Satish: Hello listeners. We are back with a brand-new episode of RawBotics. We have a unique guest with us today, someone who is a Gujrati by birth and a techie at heart. He loves solving problems that require multi-disciplinary innovation. And we all know robotics is a field which requires a multi-disciplinary approach. Please join me in welcoming Mr. Yatin Varacchia. The man who is inventing a robot to solve our cooking woes, Nosh the robotic chef of the future. So, Yatin, welcome to our show.
Yatin: Thanks for having me and I am really excited to be part of this podcast.
Satish: So, Yatin, tell us something about your journey and how do you gravitate towards robotics?
Yatin: When I grew up, I had a very high love for physics. I used to enjoy physics and how things work. So, during my school days, early college days, physics was what I was more attracted to. And I wanted to do something with that knowledge, like having that physics knowledge. So, in fact when I was in 8th-class, one thing is like India should have Maglev train. That was my dream, like why everywhere else it exists and not in India.
Satish: So, Maglev train is the bullet train, which we are finally going to get.
Yatin: Yeah, I used to read all through magazines. Like that time no internet, right? We used to read through magazines, we were imagining, I used to stay very close, where the railway line passed and they were like, why our trains are like this and not like, how it is in Japan, So, I liked the application of the engine, the sciences and then try to choose engineering. In fact, I come from South Gujarat where everyone takes pharmacy or chemical engineering, because it is actually the manufacturing hub for chemicals. So, but I chose this path because I felt like I can use this knowledge and build something on top. So, my, the high love of engineering when I joined college, I did a lot of projects like simple as metal detector which is not like a big cage, but very small and very effective. Then we built bunch of robots, photo robot phineas and robotics competition. And yeah, then I joined my master in Indian Institute of Science at Bangalore. This is the first time I came out of Gujarat and actually it was a great exposure because you are studying where Kalam and Sarabhai and everyone had studied and it actually pushed my interest to engineering very high, and then we had a department which was established by Swiss professors like Swiss engineering. So, when this department started, most of professors were from Switzerland and they were mostly into electromechanical stuff like making very small watches effective high engineering? So, I kind of realised how strong that is. And we got taught like this is a one programme which was so multi-disciplinary. It started with industrial design, a chip design, power supply design, you have problem solving. So, I worked on a problem saying 2010, where the tiger population was really on decline, and we made a camera trap which is like for counting tigers. It’s a motion trigger camera where it captures the photo of the wildlife, and it is placed in the forest. But the key thing was it should last for about a month on the battery. So, typical camera you can’t last so long. And if you make it very low power it doesn’t start very early, like it takes time to start. So, all those practical problems we solved, we deployed something in the forest and then we got to know how well it is capturing the tigers.
Satish: So, you were actually merging hardware and software.
Yatin: Yeah, so embedded software and the hardware is like a forte I like to play. That is something I enjoy because you can solve the problem only through that. Like major is the problem you can solve if you know hardware and embedded software. That lot of problem you can solve with that. So, that’s where we were working. Yeah.
Satish: So, how did this cooking robot happen?
Yatin: Yeah, we both me and my wife both were working, and we are Gujju to who are settle in Bangalore. So, we are not getting the food that we want.
Satish: So, you both were sucked in by the Bangalore charm?
Yatin: Yeah, Bangalore is really good, we enjoy here. But food is always problem because you get a cook, you train them, the moment you train them they leave. So, and most of cook doesn’t get trained. Like and we need bit of sugar in our dishes, in fact if you don’t ingest sugar, this is you either put more or less weight. That’s why we couldn’t get the taste that we wanted and that’s why I started like being from technical background. Can I solve this problem, right? This is a big problem and almost everyone who migrated to a new city faced that. So, I talked to a lot of my friends who are living either in Bangalore or in US and everyone had that complaint saying that ‘cook ka kahan pasand ni h but kya kare’, what is alternative. So, I felt it is a strong problem to solve if we can solve it. So, I started working on part-time basis and we were also looking at what food problem to solve. So, shall we make rotis and we make subji, dal or what. So, there our thought process, like there should be something where creators can build on it. So, it should not be one of the solution but it is a solution on which people can create their own stuff, like that will give mod age, it can create a mode in a business term like saying that you have so many recipes, so many creators, no one can build that kind of thing. So, that process we chose this and what we believe like this is a taste problem we are solving. So, the taste-defining thing, this is your subji and dal is what we should solve for and that is where we stared Nosh.
Satish: So, you say that you are solving for the taste, and we all know that cooking is a fine art which requires a fine mix of culinary skills and I think all our mothers were the masters of this art, right? So, how would a robot solve this problem? How is your robot or how is Nosh solving or mastering this art?
Yatin: So, actually there is a saying. What you don’t understand, you call it art. So, cooking is art. But at the same time, (So I don’t know how to cook, so I call it an art). Yeah, but it is actually a science, right? It is actually chemistry. When you cook food, chemical reacts, the enzyme breaks the protein bonds and creates something amazing. So, it is a science because it is very hard to understand it 100%. So, in fact if you see the cooking process, typically it is… You can heat and you can add right ingredient at right time, like that kind of summarises it. (You’ve summarised it very simply) Very simply. Now, the key thing is the judgement and judgement is about what thing has to go when. Which ingredient can go together and not to go together. Then, actually, lot of food science cum chemistry behind it. So, majority of cooking, if you see, it is about playing with moisture, which is you remove the moisture from some ingredient and then let the taste go into the oil, like that oil dissolves all the taste particles of it, and then you coat it to some other ingredient, which is nutrition rich. Like, paneer or something like that. So, that is briefly how the process works. How we have done it? So, we have a heating, mixing and adding ingredients, Ingredients could be three types. It could be liquid, it could be solid, and it could be condiment. Now, condiment or spice is actually defined the taste of the dish, basically in engine context. So, we have a separate spice dispenser, and this has to be very precise. So, we took lot of iteration to get it right. It dispensed the spices precisely, and it works across all the spices. That is the beauty of it. Then we have an ingredient dispenser based on the recipe, you load the ingredient, and then ingredient get dispensed. (So, this ingredient could be chopped onions, chopped potatoes, chopped tomato, whatever we require.) Yeah, it could be meat, it could be peas, it could be anything, right? So, it gets dispensed as per the recipe, as per the time, and then there is a stirring. So, when human do it, we feel like, okay, it is the stirring. But stirring is an art, because different type of food, different type of consistency, require different type of stirring, and to get it right to robotics, like one mechanism that can do stirring for all kinds, is actually very challenging. We also realise after lot of iteration, so you have industrial robot, which kind of mixes, like you will find the robot in Guru-Dwara, which mixes it, but it is good, it is designed for one kind of food, here we are talking about wide variety of range of dishes that is a (hundreds of dishes). It is created. So, that those are the key hardware part that come into picture to cook it. Second part is actually judgement, and judgements are very, very hard. So, we have camera, temperature array, which look at the pan in a different way, and then make a judgement. It can make judgement like your onion is translucent or golden brown. Your curries cook properly, let’s say you are cooking tomato. So, tomato has caramelised properly or not, And then…
Satish: So, it takes images and fixed intervals?
Yatin: Yeah. It takes images that fix intervals and temperature. So, you can cook at fixed temperature, you can cook at fixed heat, And to basically… So, in the robot, right? Why we feel cooking is art? Because we don’t know all the science behind it. So, when we have science, when our model says that put it on a sim and then stir continuously, so what she is doing is like constant temperature cooking. So, once you understand the basic science behind it, it’s actually quite beautiful. (So, how does a robot decide that how much stirring it has to do, or how much temperature it has to go to?) So, actually that lies on creator. (Okay, the recipe). So, how the hardware and software or how the product is building, on this product, there is a recipe to be written. So, it follows the recipe command. Assume this is like an Android or Apple iOS. So, it has the iOS has an Apple hardware plus the operating system. And on top, someone can build app. These are almost like that. So, you have a hardware, you have a let’s say, culinary OS. And then people can create a recipe. So, it follows the recipe and cook it. And it gives the nicer control to the creator, saying that you can cook at constant temperature, you can cook at constant heat.
Satish: So, it will give us certain set of options to choose form. (Yeah.) And you can create your own recipe. (Yes). That’s interesting. That’s very interesting. So, when you were building this robot, what was the major challenges that you felt you were trying to solve for the food industry?
Yatin: So, in fact, when we started, there was no reference. Like there is no one has built a cooking robot like this. And we had no reference So, we faced lot of problem first-hand. I think the first problem we struggled with is spice dispensing. It’s required accuracy and all spices behave so differently. So, some spices, like cumin mustard or little, small seed, get stuck in the mechanism. Anything you make mechanism. And then the separate kind, like, say, turmeric, right? So, turmeric doesn’t flow because it has very high wander walls. It doesn’t flow. Like, you need to push it. So, having a one mechanism that combined all type of spices was a big challenge we faced. Obviously, stirring was also a similar way, like having one stirrer which can solve other problems. After our first prototype is built, we had all these problems solved.
Then our second level problem started, which is about longevity and reliability. So, we realised that oil, let’s say oil is pump, and then oil over time polymerised in the pump. So, the pump performance changes over time, and you can’t have that. So, you need to have a solution which will make sure that it doesn’t happen. And you need the right pump that makes sure that doesn’t happen. Also, in food, when hardware go through food. So, food is acid, Bases, and salt all together, and then there is another layer which is coming as detergent. So, detergent is actually designed to remove stuff. And if you make a material, if food is touching something, make sure that detergent doesn’t remove it. So, that is a very complex because the life is compromised when detergent is used. So, these are the some of the problems. And then oil is another very funny part because, so we have this reference like FDA 21 CFR. So, any food grade part you build, you follow this FDA 21 CFR standard. And we realised some of the material in FDA 21 CFR which are listed still react with oil. Like, it kind of react and create a different chemical altogether. So, we went to lot of learning around the oil to get oil working. Like, oil feels very basic, very humble. But getting the oil working, basically our favourite mustard oil working (which is what a lot of Indians use) is a real challenge. Yeah.
Satish: True. So, since we are talking about tastes, it is said that the Indian taste and Indian cuisine change in every 100 kilometres. So, how are these cooking robots customised to the needs of Indian audience?
Yatin: Yeah. So, in fact, I would say taste changes from house to house. Like, you go to your neighbour’s house, The taste is different. So, how we are dealing with it. So, actually, as a company, it’s good that it is diverse so that we can create more recipe, more creator and larger mode. That’s how we look at it. Also, you have customisation. Because it’s not that you buy a machine, and you have a recipe as per your taste. So, you can customise oil level, spice level and salt level. So, that is a basic customisation that is given. So, any time you can customise it.
Satish: And all of this would happen to the interface of the app. (Through the app). You only need to write.
Yatin: Yeah. So, mobile app, you just select, like, there is a level of it. You want low, high, medium type. Okay. (So, from fixed selected options). Yeah. Fixed selected option, you can choose it and it will do that. This now is available recipe wise also. So, recipe wise also, you can customise your consistency, your spice, salt, oil, fat level, oil basically fat level and dullness. So, those all customisation, you can do recipe wise here. So, that will address the house-to-house variation. And 200-kilometre range, obviously, it will be solved through lot of recipes, (so you’ll create an ecosystem of recipes and upload it somewhere) which is like creating. So, that’s how it will get solved. Eventually, but our current thing is most of urban engines are trained to eat someone else food. It’s just that it has to be good food. So, we travel, we have seen multiple cuisines. So, we are not stuck to one food, we understand the broadness of it, and we actually enjoy that more or less. We enjoy the multi-cuisine, we enjoy different kinds of food, different kinds of culture. So, right away, we are building our own recipe base, which will be, around launch, it will be around 200 recipes, which are solid recipe, which you will enjoy as an urban engine, and then we’ll go into creator mode, where they will create more and more recipes.
Satish: Very true, and typical Indian breakfast could either be a Dosa, an Idly, a Vada Pao, Missile Pao, or an Aloo paratha, or a Poha, and all of this belong to different regions.
Yatin: Yeah, and it’s broadly, like, you are in the Delhi or Bangalore, it is the same. It’s the same, yeah.
Satish: Broadly, it is the same, very true. So, can we visualise Nosh, or the robot chef, taking over the restaurants in the near future?
Yatin: So, Nosh is not made for restaurant, but actually, there are a bunch of robots that are coming for restaurant. In fact, we also, almost two years, we have spent on B2B space, understanding how we can automate cloud kitchens, etc. So, it’s very true, it’s going to come. So, most of the branded restaurant, right now, having a cold-chain process, (so what do you mean by a cold-chain process?) Cold-chain process is like food is made in a factory, and then packed, and delivered to the outlet. Outlet does assembly and reheating, and delivered food, right? So, typical McDonald’s.
Satish: In a McDonald’s you are assembling the food, just like a car assembly and even cloud kitchens are following this.
Yatin: Cloud kitchen also does that. So, now, in B2B space, right now, that big opportunity is food assembly robot, because assembly process also changes. So, you are doing burger, you are doing pizza, you are doing chicken, you are doing pot-based meals, tacos. The assembly process is different. So, you can have different different type of robots. I feel that there is a huge opportunity there in the western world, because like people eat out so high. They mostly eat out, and then that’s where, I think restaurant is like 20-30% of their GDP. (Food away from home category is big there) Yeah, so, that is where I think opportunity lies. We were doing this in COVID time, we didn’t explore the overseas market, but there is a big opportunity there, I see that next five years, there will be lot of automation, like most of the outlet will be done by one person or two persons.
Satish: And then there will be a robot to assist them. So, NOSH will not be that kind of robot, that’s a different kind of automation in robot. So, how does NOSH look like?
Yatin: So, NOSH looked like an oven, it’s like a microwave-oven size device, where you can, spice stage inside the device. Water and oil also, it has a built-in storage. And you load the ingredient tray, where you have put all the chopped ingredients as per the app actually shows that how you have to place it. You load it, command it to cook and just tap out. You can go to gym, you can do yoga, you can watch TV, or you can spend time with your kids. (You can Netflix and chill and the food will be made). Yeah. So, you don’t have to worry about how it is made. So, and it’s very, it’s a humble looking, but very technically advanced. We have made it humble looking because we Indian like humble things,
Satish: And it has to be in the kitchen. (Yeah). So, it should look attractive. Yeah. So, will NOSH solve our maid troubles?
Yatin: Definitely, right? In fact, this invention came from my own struggle with cooks, right? Like I can’t find a cook who can cook as per my taste and doesn’t stay. So, I train him, he go away and then I will get another one. So, that’s how it is going to solve. It will require little bit effort from your side, but your taste problem will get solved. (And you will get warm home cooked food), warm, fresh, you don’t have to like cook in some time in morning, 5am, 6am and then consume in the night. This is something you will never…
Satish: When the need comes, he or she will cook at a certain time and then you eat it at a later point of the day.
Yatin: And time alignment is a big problem. Typical people, like they have to eat early, they have to wake up early morning just to open the door for the maids.
Satish: So, I can buy a Nosh and my mother’s worries that I eat outside will probably be solved. Yeah. Because I’ll get good food.
Yatin: In fact, like we have been doing trials, there are people whose online ordering became zero when NOSH was at their place. They used to spend 20,000 rupees per month on online ordering. When Nosh went in, they didn’t order, they just consumed from NOSH.
Satish: Yeah, they were probably ordering about a lot of groceries then.
Yatin: Yeah, so we have a meal kit option, so you get a chopped ingredient as well from us, so they were consuming that. Which is also a business for us, like consumables.
Satish: Okay, so consumables in terms of the chopped ingredients is also the sort of the business line for you.
Yatin: We are exploding that, shall we do it, shall we find a partner, what is the right way to do it?
But for the trial, (you could also have a robot which could chop all the ingredients). Yeah, in fact, chopping, so let’s say I said I want onion chopping machine. There are hundreds of them, all. So, as I said, in food industry, automation have existed. It is about one thing and not about everything. So, there are lot of… (There are special purpose machines). Yeah, there are lot of special purpose machines.
Satish: So, after listening to all this, I have written very hungry by the way. So, we’ll probably have something cooked from Nosh. So, can we visualise our robot chef NOSH competing in Master Chef Australia? Because I think a lot of us watch MasterChef Australia.
Yatin: Yeah, definitely, and it will have very high advantage over humans. I’ll say why. (Unfair advantage) Unfair advantage over humans, because, see, right now with ChatGPT and a lot of data being available, we have, so many recipes over internet. And this robot can learn from all those recipes. Which human can even not comprehend so much knowledge. Once they learn it, then they can make anything. So, that way it has very high unfair advantage over human, I would say, under recipe creation. And we’ll reach there, someday it will be better than human chef.
Satish: For sure. And probably there will be more and since in the recipe ecosystem, there will be something… There will be a magic that will come out from human and robot collaboration.
Yatin: So, I would say in the robotics. Next 10 year is a collaboration period almost on everything. Even we talk about ChatGPT or anything. It requires a lot of input from us to get something done nicely. So, these 10 years are going to be that period. And then, slowly, we will get freedom from this kind of tool completely. So, that’s what I feel.
Satish: So, we’ll move towards AGI from AI. So, which dish cooked by NOSH, what chef is your favourite?
Yatin: Gajar Halwa. (Oh, wow). There is storey behind it. So, when the Nosh first alpha prototype got ready, it was this time, somewhere around January. And then, we had a chef who made this carrot halwa.
Satish: And which is the favourite recipe of NOSH, like NOSH has cooked?
Yatin: That’s, like, NOSH is not that judgemental yet. Thanks to us.
Satish: So, the job of being judgemental is still delegated to humans.
Yatin: Yes. So, right now, it does what we say. It doesn’t judge.
Satish: So, all the recipes are its favourite? Yeah. Great, great. So, thank you, thank you for coming on us to sharing your experiences with our listeners. It was really delightful to know your experience and know the capabilities of NOSH – the robot chef.
Yatin: Thank you for having me. And then, I really like what you are doing, creating this ecosystem of robotic enthusiasts and, we all coming together. Like, there is no other platform where all robotics, founder, companies can come together. So, thanks a lot for having this initiative.
Satish: Thank you. So, listeners, I hope you enjoyed this episode as much as I did. Like and subscribe to continue on this robotics journey with us. Until next time, this is your host Satish Shukla, signing off!