Earlier this month, Starkey released its Omega AI hearing aid platform, and at EUHA 2025, Chief Technology Officer Achin Bhowmik discussed how the company’s latest generation builds on prior AI-driven models with new neural-network architecture designed to improve both hearing performance and health monitoring.
Omega AI introduces what Starkey describes as the industry’s first use of three deep neural networks running in parallel. Each network performs a distinct function: one enhances speech understanding in noise, another predicts and proactively adjusts directionality, and a third monitors the acoustic environment to improve spatial awareness.
According to Bhowmik, this approach—modeled on how multiple regions of the auditory cortex process sound—has resulted in measurable gains, including a reported 28 percent improvement in speech recognition accuracy and 8 decibel greater audibility for off-axis sounds compared with the previous Edge AI platform. Beyond hearing performance, Omega AI expands Starkey’s health-and-wellness capabilities. The devices now guide users through AI-based balance exercises, leveraging built-in motion sensors that assess gait and stability using protocols developed with academic partners. The system can also track physiological data such as respiration rate, providing users and clinicians with additional health insights that can support fall prevention and overall well-being.
Bhowmik also outlined new generative-AI tools integrated across Starkey’s ecosystem. The TeleHear AI feature enables wearers to describe listening difficulties in natural language, prompting the device to suggest and apply fitting adjustments that can later be reviewed by the professional.
Full Episode Transcript
Speaker 1: Hello, everyone, and welcome to This Week in Hearing at the Starkey booth live at EUHA. I’ll be talking with Achin Bhowmik. Good luck. It’s a very lively booth setting, and it’s gonna be awfully loud. Hopefully we can find a quiet spot to meet. I’m pleased to have Achin Bhowmik with me. We’re gonna talk about the new Omega AI and all the improvements made since the Edge.
Speaker 2: It’s great to see
Speaker 1: joining me.
Speaker 2: Thank you.
Speaker 1: So, I understand there are a couple of different things we need to talk about. One is how you’ve improved the DNN, and the new functionality you’ve added to the DNN. Would you please explain that?
Speaker 2: Sure. And in fact, if you allow me, I’ll take a step back for a minute.
Speaker 1: Please.
Speaker 2: You can recall our conversation last time, in fact, maybe we even discussed
Speaker 1: Yeah, correct.
Speaker 2: … after I joined Starkey, that we look at our steadfast technology direction for advancing hearing aids with three important areas. Not equally weighted, but important. The first is help people hear better. That’s what people buy hearing aids for. It’s our job number one. So that’s where the advances in deep neural network and other algorithms come in for making the products better and better.
Speaker 1: And as a person who has difficulty hearing in noise, I appreciate all the work that’s going on.
Speaker 2: It’s great. And we’ll talk about that. And then the second part is health. help you be healthy. Track and monitor your physical health, your cognitive interactions, to make you a better human being in terms of being healthy. Second area. The third area is invest in AI so that the device in your ear becomes the best conduit to a personal assistant that helps translate languages, answers your questions. The dream of human/computer interaction when you’re simply using natural language to have a conversation. So those are the three areas that Omega AI is step up from our prior generation product is, but I of course will want to start with the exciting progress we made with deep neural network.
Speaker 1: Yes, and since if you name those in
Speaker 2: Right.
Speaker 1: audiological improvement is your top priority.
Speaker 2: number one. Right.
Speaker 1: what have you done
Speaker 2: Yeah.
Speaker 1: … the Edge in Omega?
Speaker 2: So the motivation for what we did, which is where I’m gonna start for your
Speaker 1: Yes.
Speaker 2: … is that us engineers found a new trick, a new way to advance the performance of hearing aids with deep neural network. It’s been a few years now, and I’m happy to see all of the industry is excited about this line of work because simply put, we are learning from nature and copying biology.
Speaker 1: Yes.
Speaker 2: The way the brain processes sensory information, whether it’s sound that we hear through the ear or the visual information that we collect with your eyes, smell, taste, touch. The way the brain processes all that, even though brain’s very complicated, the more we learn about it, the more there is to learn. But now thanks to neuroscience, we know the fundamental computational architecture and that is neural network. The brain is a giant neural network. In fact, it’s not one neural network. It’s a amalgamation or combination of different neural networks that process different types of information. And it is not just a visual cortex focusing on the visual world. We say auditory cortex is for sound, but when you look deeper into the way auditory cortex functions, different parts of auditory cortex have neural networks that perform different tasks. And so I’m going to build on that a little bit before I talk about so what are we doing with Omega AI for neural networks.
Speaker 1: Please do, and, and I have to preface this by saying this is exciting to me because I feel like all the acoustic techniques have been exhausted. You can do binaural beamforming, you can do acoustic noise filtering.
Speaker 2: Yeah.
Speaker 1: If you’re gonna get further improvements, it has to be done along these lines.
Speaker 2: Yes. Yeah. And all that is needed, we need amazing preprocessing of the signals. We need good transducers. We need great directionality system. And on top of that, we need the next step to borrow from the way brain processes signal to help us understand information better. And again, in neuroscience, auditory neuroscience, we say you don’t hear with your ear, you hear with your brain, because ear is simply sending neural pulses and its job is done. And it’s in the amazing computational architecture of the cortex where we are understanding speech, suppressing noise, helping me be aware of where I pay attention to. There might be sound over there and I’m going there. So first, with Genesis AI and then Edge AI, we attacked speech and noise. That’s the most important part.
Speaker 1: the most important thing.
Speaker 2: That, you know, I want to understand speech better despite the cacophony of noise. So I think with the deep neural network-based speech enhancement and noise reduction, we are getting very good data, very good results.
Speaker 1: Mm-hmm.
Speaker 2: But then you look at what else is important. Well, I go back to the brain again. The auditory cortex is not one monolithic neural network. Yes, there is a, an, area in the auditory cortex whose sole function is to help me understand speech in noisy environment. How about the other parts of the auditory cortex? What are they doing? Well, we know that there is a central component to directionality. sort of like it’s not simply a if this, then that conditional statement that should drive the directional system of a hearing aid. So, for the first time, we have deployed a deep neural network, a dedicated one, that’s running in parallel with the speech-in-noise enhancement. But this one, this deep neural network, is doing a predictive step towards proactively driving my directionality system.
Speaker 1: Okay.
Speaker 2: Why is that important?
Speaker 1: you’re running two DNNs in parallel.
Speaker 2: We’re actually running all three.
Speaker 1: Okay.
Speaker 2: So, now I’m talking about the second one. I’ll talk about the third one soon. the deep neural network power deep neural directionality system, what we found in the lab years ago, and of course now it’s in the product, is that it provides much better performance in driving the directionality system in a responsive way so that you’re not constantly chasing the condition. If you do, if this condition is met, then do this. By the time you change the directionality system to that, it changed already. Now you’re changing the next one. So, you’re constantly chasing something that’s changed already because of the time-varying nature of acoustic waves. With this new deep neural network driven directionality, it’s predictive in ways that its response, its responses are not subject to the same fluctuations that you typically see with directional systems.
Speaker 1: Okay. And so then, would I perceive this as a greater sense of directionality than…
Speaker 2: You will get better speech recognition with the directionality system. If you design your directional system and measure it in ways that your directional system is fixed now and you’re measuring its performance, versus you out in the wild and it needs to be so responsive for me to understand speech in real world. We measure, we are measuring 28% better speech recognition accuracy with Omega AI with a directionality system compared to Edge AI.
Speaker 1: So, I’m assuming that’s in different directions, right? So, is Edge and Omega essentially the same if I’m facing forward, but if there are different people talking in different directions, I’m gonna have better speech recognition of all those different directions because you’re able to interactively follow?
Speaker 2: So, because you’re now adding two things. One is the speech enhancement-in-noise, that’s there. The other neural network now is doing much more proactive direction for the directionality system. So, getting those two added up. And the third thing that I’ll talk about is we have another neural network running in parallel. So, this is industry’s first concurrent multiple deep neural networks running in parallel. The third one’s job is to look out for signals in my 360 degrees surrounding environment to make me better spatially aware. So, the way I explain this, even with our biological hearing system, if you and I are so focused on having a conversation and someone calls you from, let’s say, 135 degrees and says something, it’s very usual for us to not get the onset of speech. And then we turn back and say, “Oh, can you repeat? What did you say?” That’s, you know, because you were not paying attention there to them. Right.
Speaker 1: So, it takes a second or two before you realize they’re talking your left.
Speaker 2: Before you lose the onset of speech. So, this was very common with traditional hearing aid systems as well, that we are so focused on the person in front of me that you would lose the onset of speech. So, we have now a concurrent deep neural network just looking out for important sound, whether it’s speech or other signals you don’t want to miss. And we’re measuring 8 decibel better speech audibility for off-axis sound, even when I’m listening to you and paying attention there.
Speaker 1: So, that 8dB, the meaning of the 8 dB is you’re paying attention to me and then somebody begins to speak over here.
Speaker 2: Right. It would have been an 8 dB
Speaker 1: Yeah.
Speaker 2: with my prior product, but with Omega AI, the audibility is higher. And
Speaker 1: It’s fast enough to pick up the
Speaker 3: …
Speaker 2: Pick it up, lose it. Because it’s doing it’s doing the processing in parallel. Older systems or many current hearing aids might need to stop this and focus there, but that’s already lost. So, you need some processing to be in… you need a different neural network. and the way the human brain works, you’ve got multiple neural networks doing different things. So, we are now sort of like trying to move on to the next phase of neural network where do more of what the brain does in the hearing aid itself.
Speaker 1: Yeah. And hence the three running in parallel with each other. And so then if I take the Edge AIs and I swap them out for Omegas in an interactive environment, I should then be able to understand speech coming from different directions much better.
Speaker 2: That’s exactly what it is.
Speaker 1: Yeah.
Speaker 2: you know I say, even though, you know, I’m a passionate engineer, but to a patient, none of this matters. The patient doesn’t care whether I have three concurrent deep neural networks or, you know classic signal processing versus AI.
Speaker 1: Yeah, they just want to hear better.
Speaker 2: Are they hearing better? Yeah. So, after fitting thousands of patients with Omega AI, many of them were upgraded from Edge AI to Omega AI. And we’re constantly hearing them, they’re hearing more, but not in an annoying way. They’re more aware of environment around them. So, I’d love to kind of fit you with
Speaker 1: That would
Speaker 2: Omega AI products.
Speaker 1: … you know, bless my wife, she’s so used to, we go out to dinner and she’s, “What are you trying today, dear?”
Speaker 2: Yes.
Speaker 1: And I’ll say, “Today, we’re doing the Edge versus Omega.”
Speaker 2: Yes.
Speaker 1: And I’ll try them in a loud restaurant … we might be talking, and then the server comes up and begin speaking. It would be, yeah, very interesting comparison.
Speaker 2: You know, one thing you recall, I might have told you before, you know, the reason I came to the industry and the opportunity that I saw, and I think we’re all collectively chasing that opportunity, that is to make us superhuman. So, you’ll see our language, marketing language these days, it’s, “Technology is so intelligent, it’s superhuman.” I have no hearing loss yet. I have normal hearing, but I can’t live without this now. These are Omega AIs. By the way, the brilliantly designed custom mold that Mr. Belos made for me
Speaker 1: Yes.
Speaker 2: for me. And when I removed it versus now, my speech understanding is much better than my normal hearing, particularly now that we’re in a noisy environment.
Speaker 1: Yeah. So, you’re running minimal gain, but you’re still getting all the benefit.
Speaker 2: I am getting the… Not the gain because I
Speaker 1: Right.
Speaker 2: I don’t need to apply gain.
Speaker 1: Right.
Speaker 2: But all of the signal processing we’re doing with deep neural network, that’s enhancing speech-in-noise. That’s proactively driving the directionality system. And now, that’s also being spatially aware. So, I’m not missing conversations around me, which I would with my normal hearing.
Speaker 1: It’s interesting. I’m really looking forward to experiencing it. And the other thing I’m looking forward to experience is your increasing path and health features.
Speaker 2: Right. Yes.
Speaker 1: Right? So, when … when I, did, when I had the … Was fitted with the Edge, I ran the, the steady protocol to evaluate my balance. I did a
Speaker 2: Yes.
Speaker 1: … video on all that.
Speaker 2: Yeah, you did. Thank you.
Speaker 1: And now you actually, I think of my own experience at the time I ran that, I was actually undergoing physical therapy for balance, and as part of that, I was given a set of exercises to at home.
Speaker 2: Yes.
Speaker 1: I had an app, and I could go through them and do the exercises at home. You’ve built that into the hearing aid haven’t you?
Speaker 2: Yes. So let me kind of draw that, draw line here so your audience will see the continuity of our strategy. So when you started working on balance, that was back in 2017.
Speaker 1: Yeah. It was fall detection first.
Speaker 2: It was fall detection first. I believe that was a low-hanging fruit.
Speaker 1: Sure.
Speaker 2: Yet a big value to patients. You know, even now, I have a, a patient in Minnesota. He’s actually an audiologist himself. He fits hearing aids, and he had fallen in a, an icy driveway in Minnesota winter,
Speaker 1: Yeah. I live in northern Illinois. I know how that works.
Speaker 2: Right. So he’d fallen unconscious. His hearing aids detected a, fall, sent alert messages to his wife and, and two sons. They took him to the hospital. Fortunately, he only had a dislocated shoulder. So I told this story to Wall Street Journal, and they wanted to talk to him, and we had to of course get his permission. So he’s also quoted in Wall Street Journal about he said, “Look, my hearing aid’s fall detection feature saved my life because I could have been unconscious there, frozen.” Right?
Speaker 1: Right. Because you could freeze like that.
Speaker 2: hearing aids saving lives.
Speaker 1: Right.
Speaker 2: Right? So, but then our vision was this. Detecting fall was a low-hanging fruit. It was not an easy technology to develop. We had to collect data, train the machine learning system so that it has no annoying false positives or false negatives. But we introduced that feature in 2018. Next, build upon from there. So we developed the feature you’re talking about of assessing your own risk of falling, following the steady protocol, stopping elderly accidents, death, and injuries.
Speaker 1: Good job.
Speaker 2: Right?
Speaker 1: could never remember the acronym.
Speaker 2: That’s a, the American CDC.
Speaker 1: Right.
Speaker 2: Center for Disease Control and Prevention. So we partnered with Stanford University, where I’m grateful to be a faculty member for many, many years. We collaborated with Stanford to get 250 patients to compare the balanced neurologists assessing risk manually of patient and our hearing aid’s ability to do that with the machine learning system. And they agreed really well. So we collaborated, collaborated to write a paper. It’s now published in Otology and Neurotology. The next step from there was once the patient identifies herself or himself at a risk of falling, can they improve with balance exercises? So with Omega AI, we introduced balance exercises. We have, as you know, inertial measurement in each of the hearing aids so they guide you through those exercises that tell you if you’re doing it correctly and And so you are able to do something about your risk, not just, you know, wait and see, “Oh, I got a high risk now.”
Speaker 1: Yeah. Or go to a, a live therapist like I did. Now, so are the, are the Is the exercise regimen interactive with the
Speaker 2: Yes.
Speaker 1: of the test?
Speaker 2: Oh. It is.
Speaker 1: depending on a person’s state of balance, they’ll be given a different set of exercises?
Speaker 2: Yes. By the way, the … So it’s, you know, three things, right? It’s strength, gait, and balance, so depending on where your issues are and then where you want to get You want to improve more, this can be very personalized and in
Speaker 1: Based on automatically in the app. In other words, reads by its steady results and then creates a regimen of exercises based on it.
Speaker 2: Right. And then you could double down and improve one particular area more. It’s quite amazing that even now, you know, we, we developed this for years now, for balance, four-stage balance when you are standing on one foot. How does the hearing aid know you’re standing on left foot or right foot?
Speaker 1: Yeah.
Speaker 2: Well, it turns out that’s the beauty of AI and machine learning. When you train the system with lots of data, it does things that you don’t think is possible.
Speaker 1: So it can actually tell which leg you’re standing
Speaker 2: Yeah.
Speaker 1: because you’re gonna lean a little bit because of your center of gravity, and you can pick that
Speaker 2: We can pick that up.
Speaker 1: and identify.
Speaker 2: The way you move, the way you stand. We are on this path of helping people not
Speaker 1: Yes.
Speaker 2: and you will see us do more things in the months and years to come.
Speaker 1: Yeah. ‘Cause I’m really interested at what point you’re able to assess balance on the fly.
Speaker 2: Yes.
Speaker 1: In other words, monitor
Speaker 2: Yeah.
Speaker 1: … over time.
Speaker 2: So you, you have studied this area. I don’t, don’t need to explain to you what the opportunities are, but you can see we are passionately
Speaker 1: Yes.
Speaker 2: … balance. It’s so important for, you know, people with hearing loss. There’s such a strong correlation between hearing loss and fall
Speaker 1: Yes.
Speaker 2: that it builds us, we can do something about it.
Speaker 1: Yes.
Speaker 2: So we feel really passionate about it.
Speaker 1: Oh. It’s really good in, in, in the respiration. So I’m assuming you’re using the hearing aid microphones to listen to respiration.
Speaker 2: So that was another one of those that we can, with machine learning, we’re able to lead it so accurately that even surprised ourselves. And so, you know, now I have … I can just simply pull up my app and I am curious about, you know, this is my health button. I go to respirator rate. I try right now. I look at today, 13 beats per minute. So, you know, I don’t need to use another device to count because my hearing aid is always monitoring me.
Speaker 1: so I’m assuming then you’re looking for, like, gross effects. In other words, if, if I start breathing fast and
Speaker 2: Than, than outside of your range.
Speaker 1: Yeah.
Speaker 2: Right? And it is an important vital data that you want to keep, but how many of us pay attention to it? Particularly our older patients for whom cardiorespiratory health is very, very important to track. Our dream here is that your hearing aid becomes your early warning system. It knows you have, might have some problems before even you know. So that’s the dream we are pursuing at every step.
Speaker 1: Well, and you’ve opened up a range of possibilities. Tell me if this sort of thing is on your roadmap or at least in your head…. you’re not only counting respiration rate, but you’re detecting when somebody coughs, and you can tell what kind of cough. If it’s a little or if it’s a deep lung cough.
Speaker 2: For sure. Yeah, yeah.
Speaker 1: Right?
Speaker 2: The biomarkers in voice.
Speaker 1: Yes.
Speaker 2: And we have actually, you know, we’ve been very public about this. We have published a paper where we have built multi-sensory devices that tuck very nicely into this custom mold, and we are able to measure heart rate.
Speaker 1: Sure. Because if you combine respiration with heart
Speaker 2: SPo2.
Speaker 1: oxygen level or heart rate variability.
Speaker 2: Temperature. The ear is the best place for measuring those.
Speaker 1: Temperature.
Speaker 2: So we have published a paper on this as well, on showing ear is the best place for health monitoring than any other place.
Speaker 1: And now you’ve got the battery life, I assume, that could support these things. You can have, for example, temperature and heart rate and still get all-day battery life because you’ve got overhead today.
Speaker 2: Very good point. Yeah. So that, that’s the other part of going back to neural network processing, you know, the brain’s the most efficient computational device we know. 20 watts of power, yet it accomplishes so much. Right? So the neural network approach of signal processing is extremely power efficient, allows us to preserve the battery life, and that’s why we also have sort of, like, been very persistent about neural network built into the same chip rather than an external accelerator, because we didn’t want to draw more power by moving data back and forth between two different chips or pay penalty with latencies. So it gives us more battery life. We’re really proud that this Omega AI devices preserves our more than 50 hours of battery life.
Speaker 1: So battery life is the same as the Edge?
Speaker 2: Right. Yep. And we can do more and
Speaker 1: Great.
Speaker 2: not, not have to give up battery life.
Speaker 1: Okay. No, I’m really excited to try it and try the new health features and compare its performance in, you know a loud and vibrant place, like the middle of your booth, right? Would be a great one to try. Anything else you want to add before
Speaker 2: Just maybe a minute on the third metric vector I talked about
Speaker 1: Please.
Speaker 2: which is making the device smart so you can have a natural language conversation with it, where we, we are, we have introduced a feature called TeleHear AI. What it is is you’re a patient. You just came home after getting fitted with your hearing aids. In a day or two, you realize everything is great, except these are some challenges you still have. so rather than having to go make another appointment and tweak along with your audiologist, your, your clinician might, with their help, because you put this in your in their hand, it’s part of the telehealth system that Starkey has, we allow the patient to just have a conversation with their hearing aids. “Hey I’m struggling a little bit.” “What are you struggling on?” “Well, you know, everything is great about the new hearing aids I got. I’m hearing much better.” “Great.” “Except, you know, I don’t really like the, when I turn on the water faucet or when the door slams shut, I, find it a little bit more annoying.” Or maybe, “I am having difficulty understanding my granddaughter.” I don’t know. You know, these are, these are the ways that we explain my problems to a human professional, the same way that you’re going to complain to your hearing aid. It’s going to analyze it with generative AI technology, and it’s going to say I think I understand what you’re trying to you know, where you’re, where you might be struggling. Why don’t you try this?” It’s going to push change in fitting.
Speaker 1: Okay.
Speaker 2: And then if you like it, you keep it. If not, you go back.
Speaker 1: So it will actually interact with the hearing aid and change the settings.
Speaker 2: Yes.
Speaker 1: Give you a chance to try it and say if you like that better or not.
Speaker 2: You go back to the original setting. Your professional knows exactly what changes were made, and they will be able to tweak it and tailor it. So we have another feature, and that’s for the professional. In the fitting software, we call it Starkey Pro Fit fitting software. There’s a lot of exciting updates to Pro Fit software I cannot get into the details of. The best fit accuracy is significantly better than our prior generation. But this feature I’m excited about is a generative AI-based help tool for the, for the professional.
Speaker 1: So you’re able
Speaker 2: They can have a conversation.
Speaker 1: … workflow in a fitting.
Speaker 2: Yes. You use natural language to talk to
Speaker 1: Okay.
Speaker 2: fitting software. So back, I to the, to where we started, the, we, you know, we are on this relentless march to help people hear better, mimic more and more of the processings that happen in the brain with deep neural network. Omega AI is a significant step up from Edge AI. And then health monitoring, we are on this relentless journey to make your hearing device be your health assistant and use generative AI for personal assistant. So I think in all of this, Omega AI is a significant jump up. Can’t wait to hear your inputs on it once it out.
Speaker 1: You know, that’s really interesting. I use Google Assistant all the time, so I appreciate having that access at any given time to be able to interact, ask questions, you know, get things done using Google Assistant. You have your own assistant that does that.
Speaker 2: Right. Yep. And then we use backend APIs. We always want to build on the best, right? So, you know, the vision for the future is when you want to have a conversation and ask a question, you shouldn’t have to pull out your device and talk to that large a
Speaker 1: Right.
Speaker 2: Because I got my devices already, they’re
Speaker 1: Yeah, me too, right?
Speaker 2: They’re nearly invisible. They’re my best conduit to the AI in the cloud.
Speaker 1: Yeah, that’s the, that really is the superpower part, right? I mean, my family are like amazed that like I don’t recommend giving yourself hearing loss like I did in order to be able to do that, but the fact, but the fact that I can interact with the world directly with my devices is actually really interesting.
Speaker 2: Yeah. You know, the dream we started with it, with reflecting on that. I don’t believe hearing aids need to be only for people who have lost their hearing. Technology we’re developing would be, would, would give people the abilities, like I am testing it myself. I don’t have hearing loss yet. Yet I’m getting better than normal hearing with my hearing aids, right? So it, we’re developing technology that hopefully people will love to wear, not because they have to wear.
Speaker 1: Yeah. No, that’s a, that’s a perfect way to end this. Everything you’re doing on the health side and on the hearing side, I’m really looking forward to trying it out, and I appreciate the direction that you’re continuing to go in to make people’s lifestyles and their quality of life even better. It’s great, Saul. Thank you very much.
Speaker 2: Thank you, Andrew. Always nice to talk to you.
Be sure to subscribe to the TWIH YouTube channel for the latest episodes each week, and follow This Week in Hearing on LinkedIn, Instagram and X.
Prefer to listen on the go? Tune into the TWIH Podcast on your favorite podcast streaming service, including Apple, Spotify, Google and more.
About
Andrew Bellavia is the Founder of AuraFuturity. He has experience in international sales, marketing, product management, and general management. Audio has been both of abiding interest and a market he served professionally in these roles. Andrew has been deeply embedded in the hearables space since the beginning and is recognized as a thought leader in the convergence of hearables and hearing health. He has been a strong advocate for hearing care innovation and accessibility, work made more personal when he faced his own hearing loss and sought treatment All these skills and experiences are brought to bear at AuraFuturity, providing go-to-market, branding, and content services to the dynamic and growing hearables and hearing health spaces.
Achin Bhowmik, Ph.D., is Chief Technology Officer and Executive Vice President of Engineering at Starkey, where he leads global research, product development, and technology strategy focused on transforming hearing aids into multifunctional health and communication devices powered by AI and advanced sensors. Prior to joining Starkey, he was Vice President and General Manager of Intel’s Perceptual Computing Group, overseeing work in computer vision, AI, and immersive technologies. A fellow of IEEE and multiple professional societies, Dr. Bhowmik also serves as an adjunct professor at Stanford and sits on several academic and industry advisory boards.
