When digital health products are discussed in major medical journals, it’s usually the names of doctors and researchers that stand out. But there’s another important story often overlooked: the engineers who turn cognitive neuroscience research into practical software for patients.
One of those engineers is Mohammad Haris Zia, a founding team engineer at Neurotone AI and one of the lead technical contributors to Lace Pro, an auditory training platform used by hearing care clinics worldwide.
A personal reason to care about sound and global hearing health
Haris’s interest in sound started with music. It became something more serious when he watched older relatives begin to withdraw from family conversations as hearing grew harder.
“You can be in the same room with someone, but no longer truly present,” Mohammad says.
That gap, between being physically there and actually following what is said, is familiar to anyone who has lived with untreated hearing difficulty. Hearing aids can make sounds louder. They do not, on their own, retrain the brain to pick speech out of a noisy room, keep up with a fast talker, or hold onto a sentence long enough to reply. That second problem, comprehension, is where Haris has spent his career.
He began that career in Pakistan, where close mentorship in large-scale software engineering is harder to come by. He closed the gap by training remotely with senior engineers at major global technology companies, who held him to the same standards they used on production systems.
“That experience shaped everything,” he says. “I learned how to think in terms of large systems long before joining a global company.”
By the time he joined Neurotone AI as a founding product engineer, he was already used to thinking about reliability, latency, and failure, not only about features. The job in front of him was not to invent a new clinical theory. It was to take existing cognitive-neuroscience protocols for auditory training and make them work on an ordinary phone.
Turning a clinic protocol into software people will finish
Auditory training is a structured set of listening exercises. Typical tasks include following speech against background noise, understanding rapid speech, holding information in working memory, filling in a missing word from context, and using visual cues from a speaker’s face. In a research setting, those tasks can be tightly controlled. On a phone they have to remain accurate enough to be useful and interesting enough that someone will come back the next day.
That translation problem is the core of Haris’s work. Clinical protocols tend to be rigid. Patients are not. He has spent much of his time on the unglamorous questions that decide whether training actually happens: how difficulty should rise, how audio should be timed, and how an exercise should feel when the listener is tired.
“The brain requires carefully calibrated levels of difficulty, timing, and audio cues,” he says. “Those elements are important for maintaining both effectiveness and engagement.”
Two further problems pushed him deeper into audio engineering.
The first was voice. Much of the speech used in digital training is synthetic. Synthetic voices are consistent and inexpensive to produce. They are also easy to tune out. For rehabilitation that is a real obstacle: if the talker does not sound like a person the listener cares about, attention drops. Haris worked on an AI voice-cloning approach intended to produce more natural, familiar speech, including voices a patient might actually recognise.
The second was tinnitus. Therapy that reduces the frequencies a patient hears as ringing, then layers masking sound around them, only works if the processing happens in real time. A delay of even a fraction of a second breaks the effect. He led the design of a custom digital signal-processing engine to do that work on a mobile device, under the ordinary constraints of battery, heat, and a consumer speaker.
Neither project is a catalogue of features. Both are examples of the same idea: hearing software has to respect the physics of sound and the limits of human attention, not only the logic of an application.
Why one version of English is not enough
Auditory training cannot be localised the way a website can. A listener in Manchester and a listener in Dallas may both speak English, but they do not hear the same vowels, the same rhythm, or the same background of everyday noise. Phonemes, accents, and speech patterns all change how the brain processes sound.
Haris treated that as an infrastructure problem rather than a translation problem. He designed an audio system that can serve region-specific sound assets, so training in the United States, the United Kingdom, and Australia can use local English without rewriting the underlying protocol. The structure of the training stays the same. The sound of it does not.
The same discipline showed up when the software had to run as a health application rather than a prototype. Healthcare software is unforgiving. Audio has to arrive on time. A session cannot drop in the middle of an exercise. Health data has to be handled with care. Haris led a shift from an early architecture to a modular, service-oriented backend, with caching, autoscaling, and the encryption and access controls that health applications require.
“Trust is essential when dealing with health data,” he notes.
Thought Leadership and Looking Ahead
Engineering is not the whole of his work. He has served as a judge at international hackathons, evaluating other people’s technical choices under time pressure. He also works with audiologists and neuroscientists, including Dr. Cliff Olson, on how auditory training reaches people who will never set foot in a research lab.
Looking ahead, his own priorities are more region-specific content, stronger infrastructure for users outside a handful of English-speaking markets, and, over time, contributing to research and engineering frameworks that other digital-health teams can use.
“At the heart of my work is a simple goal: helping people feel connected,” he says. “I’m drawn to tools that make it easier to hear, understand, and truly be part of the moment.”
