Qeexo develops machine learning solutions for highly constrained environments. Small, but powerful, Qeexo’s machine learning engines are lightweight, delivering high performance with an incredibly small footprint. This allows its models to run locally on devices without having to go to the cloud, making it ideal for low-latency applications in mobile, IoT, wearables, and automotive.
Transcription:
I am very excited about this next guessing of Audrey because you know, we were just talking to him last week. I didn't realize these guys will be on the show again.
I know he's a repeat. Okay.
I think it's awesome though I am so pumped. I'm like, Wait a second, guys running a little company here called Qeexo that's working on some really cool AI engines. And I get kind of nerdy about that kind of stuff. You know, I don't know much about it. But I'm always just completely like entranced. So we have Chris Harrison back on the show. And he's not like rebuilding computers from 50 years ago. He's also running a really cool AI company, Pittsburgh, Chris.
I would call backwards forwards guy.
Yeah, you got to look both ways if you want to run in the future.
Mm hmm. Wait. So tell us about this company. And are you and what's your role in the company? Are you the founder,
so I'm a co founder and I'm CTO. And our CEO and the other co founder heads up the office that's in California. So we have a kind of a split HQ.
So tell us about the company, give us the elevator pitch and then we'll jump in over.
What are you solving?
Yeah. So Qeexo started as a machine learning solutions provider for the mobile business. So smartphones and things like that embedded systems edge computing. And we shipped a lot of different units, different kind of products that were more consumer facing that we showed at CES one a bunch of awards at CES. And then over time, we realized actually, that our technology stack that we've been using to make these very advanced kind of AI engines for mobile devices, which is a very special context, very constrained in terms of compute, they don't have a GPU. And we turned, we actually found out that that product, our internal tools was actually really powerful. So we've launched a new product called auto ml, at the end of last year that we started to give to customers, they can use our internal tools now as their professional development program.
So is this like, would you say you're going to be key so inside
Maybe so we do I mean, on some on certainly on some phones, you can actually go into the about, like Pedro in the like a phone information page and you'll see our logo in there. If it's running our software now we are running on something on the order of 300 and 50 million smartphones right now.
What are you serious? Really? Wow.
Are you taking investment money right now?
No, not really. I mean, we're always happy for people to knock on our door, but we've got some really record quarters recently. And so we're, you know, we're profitable. With our Alright, well, then forget it, Jonathan, and I aren't going to give you any money.
We're good, but now we don't have to. So how many how, um, let's think about this. So you have half of your company on the west coast. The other part here? How many years in Asia as well?
Okay, so how many people in Pittsburgh? And are you looking for talent?
Yeah, we're actually hiring right now bringing on a lot of people right now. We have ads out We're probably about a headcount of 20. Right now, I haven't been in the office in so long, we have quite a number of interns as well. I can't see everyone in the one zoom screen is all I know you have to be different.
And so who are your customers? So who are you selling to right now?
So in the early days, we were selling to the big mobile phone OEMs. Right. Those are the sort of Samsung's of the world. Right? Now, we're increasingly selling our core technology, because again, when we cutting our teeth in the mobile industry was really interesting, because they said, Oh, hey, we want to apply machine learning to touchscreen data. But we only give you one millisecond of compute, right? And it has to, you know, take this little amount of battery, you can only run in 500 kilobytes of RAM. We're like, how do you run advanced AI in such a small I mean, that's ridiculous. Yeah. And while everyone else is going to the cloud, right, then we don't go to the clinic and go to the cloud for a touch event. Right? Your machine right? Do you want to run on the machine and we don't want to Slow down the touchscreen. So it has to be really fast. And we're like, well, that's pretty tough. But by virtue of doing that, we got really good at taking sort of advances in AI and packing them down to run really fast and really lightweight on these small processors. And so by building that out, it took us two years for our first commercial product, and then that went anywhere. And then now we've been live for, you know, six or something years at this point. And we've got this really amazing platform. And so now we've released that as its own standalone product, if anyone's developing machine learning AI, that has to live on a small processor, not a big GPU that's on some gigantic cloud farm, you know, sitting in some warehouse, but has to run on that little tiny device, that consumer device that might only cost 10 2050, maybe $100, like a thermostat, a toaster, whatever you want, you want to put Martin's module in a car, something like that, right? You're gonna turn to development tools like what we're offering in this auto ml platform. And so it's really just sort of a consumer ification of our internal development suite that we've been using for five years and now our new customers are not so much Yeah, I was gonna use it too, but it's more of the chip makers or the sensors, makers, you know, someone's making an accelerometer and they say, we want to do step counting or we want to do are you running? Are you bicycling? Well, that requires machine learning to live inside that little $2 accelerometer how do you do that? That's a really tiny computer That's so crazy and you guys make that happen that soon it's like the bulk your engineering happening here in Pittsburgh?
Yes, our main kind of r&d facilities and all the machine learning you know kind of talent is here for goodness sakes awesome.
So when you when you're selling your product now right, how are people finding out about you? So what what's that strategy? We got?
Yeah, I mean, we are people who are in that industry know about us. So when we a lot of the big kind of events that happen in that space you know, trade shows, you know, arm had their big arm Connie or all the vendors that use ARM processors come with just pretty much everyone and they wanted us to come and present on their own floor. Rhode Island thing. So we getting an increasing number of invites to show off the technology working with partners, we have partnerships with people like Qualcomm, I mean, all that kind of heavy hitters are doing machine learning. They're inviting us to participate and have kind of formal partnerships with us boss as well. So, so that is good visibility. But also at the same time, our new product is fairly neat, you know, before we sort of lived kind of under the carpet, you know, inside of someone's mobile phone, no one really knew that we were gay, they went through the settings. And now we're sort of a development tool. So it isn't really consumer facing. So we're not like a name brand that a consumer would know.
Right? Right. Right. So but so what about COVID? Has COVID affected your work your performance, either positively or not?
I mean, it has I don't think it's helped our productivity in terms of a company but it has opened up interesting use cases and you know, a lot of customers come knocking on door saying, hey, we've got some sort of door sensor, you know, you look at what's happening in a market. They want to count on it. Yeah, sure you've been to playing $500,000, you know, face recognition system know that, but that takes time takes money. So a lot of people are looking to add that sort of little dusting of magic onto exam infrastructure. And you know, you can do that with other machine learning. So we are getting, you know, it's like a business, we really want you I think we'd all not prefer this not right now, but it has opened up some interesting business avenues.
And so what about the kinds of people that work for you? What are some of the skill sets? And what are you looking for, like in terms of expansion?
So in the Pittsburgh Office specifically, we have a really awesome machine learning team, and they come from very different backgrounds. We have the kind of more traditional computer scientists, maybe even someone who took a master's degree at CMU in machine learning, but we also have economists. We have roboticists, we have physicists, we have neuroscientists, and the list goes on. And what's really interesting about it, actually, the team lunches are awesome. Because we go into like, really mystery talk a lot about technology talk a lot about biology. But what it means is that each of the different fields like, you know, what an economist uses to understand data, what a neuroscientist is understanding are quite different. They have different tools, and all, you know, often has a similar mathematical kind of, you know, Foundation, but what kind of tools they can pull out their toolbox of skills to attack a problem when someone comes to us with an interesting question is awesome. So it's a great team to work on, because it's really an incredible depth.
And so what do you think like, you know, people are probably working from home now, that's probably, what do you think for the next six months? Are you going to bring people back to the office? Or do you think that you've been able to use this interdisciplinary model through the virtual world effectively?
Wow, that's a really tough thing. And I'm also not like, you know, in every meeting on a day to day basis with the engineers are going to extrapolate. I would say that, you know, we are cautious. I think we actually closed the Our office here before really, you know, it was trying to do we did that out of an abundance of caution. I think we there was a desire to, to open the offices, but I think we're going to follow sort of the science on this and follow the guidelines. I think we I think we were on track, like many people were that when we saw things were improving maybe a month ago, like always sort of this glide path back to some sort of degree of malady. And now it's sort of hit this, you know, hockey stick, which no one really expected people did expected, unfortunately, but it's worse than we might have imagined. So I think it's gonna be a while before we really have any concrete plans of re inhabiting the office. But certainly everyone likes to again, it's a really collaborative atmosphere. We're all sitting in a shared space. So there's a lot of discussion, right? And we do miss that all of us miss that.
But there are probably ways that you can continue to do that. So what positions are you recruiting for right now,
mostly, it's software engineering, and machine learning and data science. Especially if you're, if you're good on sort of the kernel or web front, I mean, two opposite ends of the spectrum really low. Really high those needs and they can see our job postings on our website qeexo.com.
And then then we can also look for if I'm were if I'm living outside of Pittsburgh, and I'm not ready to move into Pittsburgh, would you still look to hire me? I think it's a case by case basis. We certainly will not turn any qualified applicant away. And so we encourage everyone to apply. And actually, I would say half the office are people that have moved to Pittsburgh, yeah. Access California, New York. So so so if it is a hub to recruit people, we have normally fly them in for on site interviews, and that's a mix right?
And where your offices when you did have offices or will go back to your part of town, you guys.
Yeah, we have great offices in East Liberty on the top floor of a building we're kind of across from Duolingo Ah, very great.
You ever throw like software's over trying to hit their windows?
Yeah, we've been building up our muscles to challenge them to a football match.
Number is a little bit right now, unfortunately.
That's great. Well, thank you, Chris. Thanks for being back. And people can go to QEX. Oh, and you can find out more about them. And they are definitely one of our rising stars right here in our region and really appreciate you building that part of the company here, Chris. Absolutely.
Story. good story. Chris. You're the best. Thanks for being part of tech vibe. Tonight. You are making Pittsburgh super proud. That is for sure. Audrey another tech vibe under our belt. It goes by fans because we're having a good time. Love telling these stories. This has been Jonathan Kersting. And this is Audrey Musa. Once again, we're from the Pittsburgh Technology Council at PGH tech.org learn about all the things the tech council does.
Transcribed by https://otter.ai