That Tech Pod
Welcome to That Tech Pod, a podcast co-hosted by Laura Milstein, Gabi Schulte and Kevin Albert. Each Tuesday, That Tech Pod will feature in depth discussions about data privacy, cybersecurity, eDiscovery, and tech innovations with heavy hitters in the industry. Subscribe so you don't miss an episode! Visit thattechpod.com for more information.
That Tech Pod
No More Language Barriers: Inside The Tech Behind The Ambassador With CEO Andrew Ochoa
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
We’re on vacation this week! While we recharge, enjoy this “best of” episode from our first year back in 2021. We’ll be back next week with a special new episode featuring PlusNXT!
Today (June 8, 2021) on That Tech Pod, Laura and Gabi sit down with Andrew Ochoa, CEO of Waverly Labs, to unpack the tech behind The Ambassador, a wearable translator built to remove language barriers. They dig into how real-time translation works, the hardware and software that power it, and where it shines in the real world, from travel and classrooms to business meetings. Andrew shares what accuracy really means, the tradeoffs between speed and quality, how the team thinks about privacy, and what might be next for multilingual communication.
Hello and welcome to that tech pod where we discuss all things e-discovery, data privacy, cybersecurity, and tech innovations. I'm Gabby Cholte, and I don't know that much about technology. And I'm Laura Melstein, and I know a little bit about technology. That's why we are bringing the experts to us. That's right, each week we're talking to heavy hitters in the industry to help us break down these topics. So today, Laura, tell us, who are we talking to?
SPEAKER_00Yes, today we are talking to Andrew O'Cola, CEO of Waverly Labs. Andrew oversees all business operations and sales as well as marketing strategies. And the reason we wanted Andrew on the show today is because of the ambassador, a really exciting and innovative technology out of Waverly Labs. It's a translation tool that allows you to have smooth conversations free from the constraints of language barriers. But before we bring Andrew onto the show today, we are actually going to test out the ambassador ourselves. So I have with us a guest today who would like to remain anonymous, and she speaks multiple languages, so we are going to utilize our skill set to test the ambassador. Okay, all right, let's try that. Okay, so you're gonna say something in French. Okay.
SPEAKER_03The view of the mountain is beautiful.
SPEAKER_00So it says in the app, so I couldn't hear it in my ear because I held it to the mic, but did you say the view of the mountain is beautiful? Oh all right. Now I'm gonna say something and then you hold up your earpiece to the mic, and we'll see if it translates in French for your mic. So you just hold the mic here, and then I will speak into my mic. I love listening to that tech pod. It's incredible.
SPEAKER_03I don't like text.
SPEAKER_00Yep, it works. Is that what? Yep. Uh it looks like it says Todd. I love listening to that text Todd. Uh guy doesn't know what a tech pod is. That's upsetting. That's upsetting. Yeah, we need to note that to him.
SPEAKER_01Um do you want Portuguese? I can't.
SPEAKER_00Yeah, let's do Portuguese a very fun one. Okay, so regular Portuguese, not Brazilian. I like that it gives you these options though. All right, so I'll speak to you first, and then you hold your mic here. We'll do an opposite order and uh so people can hear what it if it works for Portuguese. I love technology and software. Does that work? Yeah. All right. All right, and now you say something, and I'll hold it and we'll see if when you speak Portuguese, if it tells the audience in English.
SPEAKER_01Okay.
SPEAKER_03Spent the day at the beach with my friends.
SPEAKER_00Spent the day at the beach with my friends. Would you use this in real life? Yes, I would. Do you think it's a cool product? I think it's awesome. I think it's awesome. I think it's a great product. Yeah. And uh I think I'll buy one. Yeah, I think it's great. I think it just needs to learn how to save that text pod all the time and it will be ideal. That was amazing. That was so cool. That was that was so cool.
SPEAKER_01All right. Well, without further ado, should we bring Andrew on? Yeah, let's do it. Andrew, welcome to the show.
SPEAKER_02Hi, thanks for having me.
SPEAKER_01Yeah, absolutely. Um, so Andrew, before we get into, you know, kind of the more tech part of it, can you just tell us sort of the origins, the mission behind Waverly Labs, and sort of the reasons behind uh you all developed The Ambassador?
SPEAKER_02Yeah, so that's a good question. I think um kind of started maybe seven or eight years ago. We were just a small team of colleagues, really, me and the original co-founders. There were four of us. Um, but we hailed from different backgrounds, different cultural backgrounds, different countries like India and Mexico. We had lived and worked um all over the world in China and Africa, and we just had these personal and professional experiences of struggling with a language barrier, both in our personal life and in our um professional environment, international teams, that kind of thing. And we were really inspired by what by what we were seeing begin to emerge in the voice and the wearable tech space. This is like in 2012, 2013, you know, things like smart watches and voice-enabled products were very much in their nose. Um and so we were just kind of collaborating, I guess, over the course of a few months about some ideas that we had, and and uh eventually we settled on the idea of like let's um let's merge these technologies and our expertise and voice and convertible tech into a product um that would solve the language barrier and how do we do that and what does it look like and you know that kind of stuff. And that really became the genesis of the idea of our first product, which was called um pilot smart earbuds, which were the first earbuds that could the first wireless earbuds that could do uh live translation, which we debuted in 2016, and then in 2018, 2019, we pivoted from the consumer space to our new product, which is Ambassador Interpreter, which is really tailored for professional um and prosumer users. And so now our vision really as a company is to develop voice and language products for professionals and teams. Um and that's really been the evolution, the evolution of the team and the company over the past seven or eight years.
SPEAKER_00That's really um cool stuff. It's really exciting stuff and and innovative for sure. Can you dig into a bit more on the actual technology behind the ambassador? Uh are you using any type of AI? How does how does that work?
SPEAKER_02Yeah, certainly. So um it's AI is very uh it's very new age, really. The the mechanisms of AI which are which hinge on neural networks and deep learning, the this step has only been in the ether for the past five years, maybe six years at best. It wasn't really until uh until about 2015 or so when the first AI-enabled services began to enter the market, and AI is what powers um a lot of the components of the entire speech translation system. So in the speech translation system, what you have is you have three um systems working in tandem. The first is speech recognition. So this is where when you're speaking, uh the your audio signal is captured and then it's transcribed into text. The second part of the engine is or the system, the second engine is the machine translation. So that happens the next step. Once we get your source, um uh the source text, the source language into text format, uh, when once it's transcribed using speech recognition, then it's translated. Um so it goes from source to target language. Uh and that's how that that's the translation part of it, the actual translation part of it happens in the middle. And then the third step is the speech synthesis or the text to speech. So now once we've transcribed, once we've transcribed it and then translated it, then we have to resynthesize it into speech so that way you can hear it. Um so those are the three main systems, and each of them are using a some form of AI in the um in the you know in in 2021. They're all using some form of AI, which is to say that they're all based off of uh all of those engines are based off of neural networks. Um uh or and and use mechanisms of deep learning and machine learning to uh improve the accuracy the more that the more they're used.
SPEAKER_01Yeah. Yeah. And when we were kind of looking into this, um, you know, we and and you said this too, that you're hoping to sort of bring this technology on sort of more professional level to teams, uh, et cetera. And I found it was sort of interesting because I hadn't thought about this before. Um, because obviously you think about, okay, when you're traveling, this is a really useful thing to have. But um, you know, different industries could find this tool really, really helpful, um, like education, like hospitality, um, and obviously the you know, travel in industry and everything. Um, so what excites you the most about that? And what's something that you think, you know, maybe we're not really thinking of that you think could be a useful thing in this in this sense?
SPEAKER_02Yeah, there's certainly there certainly is a lot of, I mean, the first when you talk about translation, like earbud translation or whatever, the first idea that pops into people's mind is, oh, I can travel anywhere and like you know, understand what's uh what people are saying around me, which is of course you know what we want. We want that to we want to uh uh appease the leisure traveler as well as much as we want to appease, as much as we want to build a solution for um for for an organization or a business. Um but it's the the the need absolutely uh is more uh surpasses the the leisure traveler. That's not the main use case. That was the use case with our first product, the pilot smart earbuds, um, which was which is more consumer focused. Um but since we've left the consumer market for a number of reasons and moved into the more professional um and consumer space, what we see now is a lot of customers come to us um from I mean, the entire they they span the entire spectrum of industry. Like uh I'll give you an example, you know, there's an OBGYN clinic in Nevada that invited us to their hospital. Um they survey, or their clinic rather, they survey um a very diverse and multicultural community, and they struggle with with uh communicating with um with their patients about what the processes are um for you know for patient care. Um another Nevada story is that there's a there's uh which is kind of funny, there's a um a very large and well-known wedding chapel in in Las Vegas, and that it's a significant portion of their of their customers from um from overseas that come to Las Vegas for a week or a weekend or whatever or whatever to get married. And so if you can imagine the staff and the videographers and photographers like struggling to bridge that language gap with the with a wedding party. Um and then of course, as you mentioned, education is a is a really big um um opportunity as well. We get a lot of we get a lot of um inquiries from school districts and and uh and uh educators who are either struggling with communicating with their students who don't speak the language natively or or struggling to communicate with the parents of the students who um in conferences who might not speak the language. So it really does span the spectrum of industry. Then you have the you also have the um use cases which you know are a bit more um uh uh evident, like travel and and hospitality and tourism and things like that. But right um it's quite a ubiquitous problem.
SPEAKER_01That's so interesting too, because um uh, you know, when you when you think about things like the medical field, and obviously, I mean, I think you know, uh Americans have a stereotype of like not let wanting to learn a lot of other languages and stuff like that. But you know, we some of us are taught, you know, second languages, but even then, like you have to really, really be deep in the training to learn like medical Spanish, you know what I mean? Um does it and sorry, my dog is mad. Yeah, um can we get that annual interpretation? Yeah, give me one give me one second, sorry, but yeah, you can answer.
SPEAKER_02Should I just say wait? Okay, no, no, you can uh so no, that's it, but that's that's a really good point. Like, you know, we call that domain-specific vocabulary, um conversation, general conversation like we're having now versus um conversation and terminology that's uh um from the medical industry or finance or legal legalese or something else. Like these are these are these can be very delicate convers um um topics of conversation or terminology. So um it it is difficult. You can have a general understanding of a language, but that doesn't necessarily mean that you can communicate across every domain.
SPEAKER_01Right, exactly.
SPEAKER_00Yeah. Um, so full disclosure, Andrew, before we brought you on today, we tested the ambassador. We wanted to see, you know, how does this work? How can we use it? Because um, personally, for my job, I'm in a global role. So I'm going to different countries, I'm talking to people that are speaking different languages. And so when I saw your product, my mind was blown. I was like, this is incredible. I'm gonna make all the sales now. It's gonna be life-changing. And so I wanted to see, you know, how well does it work? So before we brought you on, we did a little test with a few different languages. We did Portuguese, Spanish, Italian, French, I don't know, just lots of different languages. And it's incredible. Um, the person who um I felt just right on speak all those languages. Um I won't share which ones I do or do not speak, if any, but uh, but the person that we had brought on doesn't have experience when it comes to technology. And so we wanted somebody who just speaks languages to see would you would she be able to use this? And it was incredible. I mean, she was blown away. She asked uh one of the questions she actually wanted us to ask you on this is how she can get stock into your company. She was that uh interested and invested in it. Um, but one thing we didn't test, which I'm thinking about right now as we're recording this, is you know, there's a global pandemic going on. Surprised we like to bring this up on almost every episode. And because of that, it's having so many impacts on things. And what I'm wondering right now, because we didn't test this, I know that this uh ambassador, I can talk, and she can talk, and you can talk, and we can have multiple conversations in multiple languages and it can speak to us, which is incredible. But can it listen? Could I use it through a Zoom conference call during the pandemic? And if not, are those some changes that you would make to the product so that way if I'm talking to you like I am right now, that we could do it? Because I I don't think you could while I'm thinking about it, since we'd have to kind of be in the same room. So are there changes with the pandemic around that?
SPEAKER_02Yeah, that's a that's a really good question. There's a lot to unpack there. Um the so it's it's it's tricky. The um the the the the system, the the product ambassador interpreter, and the sit and the and the system they're in is really designed for face-to-face communication. So if you and your guest or you and um you know your your business colleague or whoever or whoever it is are in the same room and you're communicating back and forth, that's not the problem. But as soon as you move into teleconferencing and you're trying to capture the audio signal coming from your computer's speaker or whatever it might be. Uh there um that presents challenges to the speech recognition system, which is the first system uh through which the entire process begins. Um typically these speech recognition systems are built off of acoustic audio, acoustic data. So me and you speaking in person. Uh and these these are this is a data set upon um which is fed into these speech recognition systems, but it doesn't do a good job at picking up uh digital signals, digital audio for some reason. So we've tested it, and uh we are playing with some models that will allow you to capture like you know, not only teleconferencing, but like listening to uh translating a movie that's that you're playing or a podcast or something like that, you know? Um these digital sources can pose an issue to the speech recognition engine for for whatever for a number of reasons. Frequency artifacts, like just there's a number of reasons why. Um we're working on a on a model that will like uh solve that problem. Um I think the best solution for us in the uh in the interim is to uh design a remote service upon which you could just allow someone else to conference into your chat room essentially um while you're having your your your your teleconferencing session. So my point is to say that like we're working on this problem. Um it's it's quite more it's more profound than just like recording audio from from your computer. Uh but we are we are very aware of what's happening. The I think the um internal consensus of the within the team is that like teleconferencing and remote work is pretty much going to be here to stay to some degree, even after the pandemic ends. And so we want to we want to um somehow uh enable the the solution to to work within those environments as well.
SPEAKER_00Yeah, I mean I think that makes sense and and I agree. I think to an extent the uh new form of communication isn't going anywhere. But I also think as things open up more and as the pandemic shifts a bit, people are going to still be doing face-to-face, and you still, when it comes to work, that happens a lot and that is necessary. And as someone who personally tested this product, uh I can't speak enough uh good words about it.
SPEAKER_02Cool, that's great. Um, yeah, glad to hear that. I'll tell the developers, they'll be happy.
SPEAKER_01Yeah, um, tech pod approved. Uh so lastly, lastly, Andrew, um I know you talked about this a little bit, kind of, you know, the the mission looking forward, but what about sort of a long-term view? Um, you know, uh like we said, obviously when it comes, you know, the first thing that people think about is kind of leisure travelers, but then you have your professional uh class. Um, so what's kind of the next step or maybe some new technology trends that are coming down the pipeline in about five years that you're like, okay, yeah, that's something that we want to really work on.
SPEAKER_02Yeah, um, I love that question. I love uh there are some big things that are happening in in the realm of voice and language that uh we we want to be at the forefront of, we want to be ahead of. So I say and I would say in the next five to seven years, here's what we're looking at. Um the first thing is we want to get there's probably three main things. The first is we want to get to offline or edge computing at the very least. Uh that is to say that right now our our entire solution is is is uh cloud-based, so you have to you can use the earbuds, but you have to have some kind of connectivity to the cloud where our servers are. Um whether it's Wi-Fi or 3G or 4G, LTE, whatever. You have to have some kind of connectivity. Uh it would be great though if you could just like use these earbuds in the middle of the Sahara Desert, right? Like um that's that's a that's a difficult problem. That's a computational problem at the like at that chip level um to solve how we can store vast amounts of of data in a server farm in Virginia or Chicago or wherever, but um to somehow squeeze it onto a chip. I can handle that that degree of computation is um more difficult. I think in I think in five I think in five to seven years, maybe even yeah, I think five in five to seven years um we could get there at least in some degree. It can it can be done even maybe three to five years. Um but going offline or going to the edge will be a major breakthrough. Uh the second one, which would be really exciting, is achieving uh what what I can what I think of as true real-time translation or what's called simultaneous translation in the uh interpretive world. So right now the way it typically works is um interpreters will interpret one of two ways, a human interpreter. Um either consecutive or simultaneous, and consecutive is where you speak and then you stop speaking, and then the interpreter says translates what you said.
SPEAKER_01Okay.
SPEAKER_02Um and so that's the way machines work right now. When you say something, the machine takes all that data processes that once you once you stop speaking, rather once the machine once when it hears what um uh you finished, basically, the machine will take what you said and then it'll translate it. But what we want to achieve is simultaneous, and that's what a human interpreter when a human interpreter uh is transiting right along raw right alongside you as you're speaking. And machines machines haven't figured that out yet. Uh it's a very difficult problem for a number of reasons. But um there are certainly Some researchers uh working on this. As a matter of fact, we have we're working on it right now. We have some world that we're developing, um, that we've been developing before we want to take that. I think we could see a beta of that within like this year, maybe within the next twelve months, at least twelve months. I mean, yeah, it could be this year even. Um but that means like once you get to that level, then it's a very sci-fi. Like it's just translating almost instantaneous.
SPEAKER_00Sci-fi is the right word. It does feel like very like surreal. Gabby, can you um help him develop this? I feel like this would be your expertise. Can you step up? Um yeah.
SPEAKER_01Oh, I was gonna say, when when you have a beta, I mean, let's let's hear it here first at that type of.
SPEAKER_00Yeah, we are big fans and we think this is one of the coolest tech innovation projects that that we've gotten to kind of dive into. So yeah, come back to our show, let us test it. Let's let's test it live. Uh, all of them, all of your beta projects, because we're we're gonna be with you throughout the whole way.
SPEAKER_02Yeah, I'll be happy to come back and and show it off. And then just to finish that thought I'll be um out with the other major innovation I think that'll happen in the next five years, I mean it will be the will be to um will be to what we call future speaker characteristics. So it's a big translator what I'm saying, but if it can sound like my voice to match my emotions and excitement, um that'll be a major innovation as well.
SPEAKER_01That's so next level. That's like cyberpunk, like 2049. I like cyber punk.
SPEAKER_00Cyberpunk. Yeah. Andrew, did you know that's what you were creating? Cyberpunk.
SPEAKER_02I haven't played the game. Uh uh.
SPEAKER_01I think we just made it up. I think Gabby just made it up. No, that's a word, Laura. That's that's a genre of film and oh it is. I thought you just made this up.
SPEAKER_00No, I uh turns out I'm not that cool. I don't know.
SPEAKER_01It's also a video game. I'm a nerd. All right. I my uh my uh my facade has fallen. I'm a nerd. Um but yeah. But that's so cool, Andrew. And you know, that that's all super exciting, I think, especially especially with the things that we were talking about, uh, you know, when we're thinking about different uses for this, not only in travel, but for the medical field, for education, for um humanitarian efforts, you know, when you're talking about taking it off um, you know, being able to use it in more remote areas. That's probably really exciting and opens up a lot of opportunities. So thank you so much for coming on the show and joining us and explaining this this awesome uh technology. So thank you so much.
SPEAKER_00Yeah, thank you, Andrew. You were amazing, and we look forward to everything that you're gonna do.
SPEAKER_02It was my pleasure. Thanks a lot.
SPEAKER_01So, Laura, we just talked to Andrew Ochoa, CEO of Waverly Labs, and he explained to us this awesome technology, the ambassador. What did tell me your take take uh tech takeaways?
SPEAKER_00I mean, this was uh probably one of my favorite. Uh I almost said this was my favorite episode, and I stopped myself because there's been so many good ones that I just don't know which one is my favorite, but I loved it. This is up there for sure. And and being uh one of our first tech innovation podcast episodes, I I thought this was such a cool product. And I was so excited when we did the initial test before we brought him on. I was a little hesitant in thinking it's not gonna understand every word. And it was incredible. It was so smooth, it was incredible. Um, and him talking about adding the edge computing technology. I mean, they're they're gonna do big things.
SPEAKER_01It that was really cool to hear kind of what's coming down the pipeline in the next five to seven years, you know, with the simultaneous translation, with getting off of their the cloud. Um, what was the third thing that he said?
SPEAKER_00I mean, I I was so enthralled by everything he said that I I I couldn't even tell you. I was just like, yes, yes, yes. All of the above. All of it, all of it.
SPEAKER_01Well, well, yeah, and that was just really cool to hear like what could be possible in the future and and also just like the different ways in which this is useful, right? Like obviously we think of like, yeah, I can go to like anywhere and not have to learn a language, or or it could like help me in my learn language learning process, right? Um so so that's cool. But then also when you think about actually, you know, in crisis times and in medical, you know, situations, that's really important to not have that language barrier.
SPEAKER_00I mean, we need to just love the pandemic end, travel the world with these, test it out in every country, that tech pod road trip again, uh part 19. I don't know how many road trips we need to do. Roadshow. Yeah, yeah. We can do it. Going on the road, everyone. We're doing it. We have to do it because we have too many cool things that need to come with us on the road. And the ambassador is definitely one of those. So if you're if you're listening to us right now, hopefully you are, and you're interested in finding information on the ambassador or waverly labs, or you just want to hear more tech innovation, reach out to us at thattechpod at gmail.com. Check us out on our website at www.thattechpod.com. Uh follow us on LinkedIn. You know where it is. Uh LinkedIn dot tech pod, I don't know. But search for us, find us, follow us, subscribe, and uh, yeah, Gabby, tell us more.
SPEAKER_01Um, well, there's a really easy way that you can help us out um and support our podcast. And it basically all you have to do is go to Apple Podcasts, give us a five-star review, and um you can leave a comment if you like, or you can just give us a five-star review and just let us know how we're doing. We love to hear your feedback, and we'll even read it out on one of our episodes. Um, so thank you so much. Maybe.
SPEAKER_00Maybe if it's good. Thank you for listening. Happy Tuesday. Happy Tuesday.