He spent a career making chips faster. Then he pointed all of it at the slowest, hardest problem in medicine: teaching a brain to relearn.
Veera Anantha runs Constant Therapy Health out of Boston, and the product he is most proud of cannot be benchmarked in gigahertz. It is a person, recovering from a stroke, who picks up a tablet at the kitchen table and finds the right exercise waiting - not too easy, not too hard, the next rung on a ladder no clinician could custom-build by hand for every patient, every day.
That precision is the whole point. Constant Therapy is a mobile app that uses artificial intelligence to help tens of thousands of people living with neurological conditions - stroke, traumatic brain injury, aphasia, learning disorders - regain the skills that quietly run a life: speaking, reading, remembering, planning. The app does not replace the speech-language pathologist or the occupational therapist. It extends them, into the long stretch of hours between appointments where most of recovery actually happens.
Anantha describes himself, plainly, as "a hands-on technology executive and business leader with a passion to bring positive change through the power of data and AI." The plain language hides an unusual resume. Before brain rehabilitation, he built things that were measured in raw speed and scale: mobile hardware at Motorola, wireless network software used worldwide, and the world's fastest digital signal processor at a startup later acquired by Apple. He knows exactly how to make a machine go fast. The interesting turn in his story is what he decided to make it go fast for.
"I wanted to really make a difference by using technology." - Veera Anantha
Most AI products are happy to be a black box. Anantha decided his could not be. The system at the core of Constant Therapy is called the NeuroPerformance Engine, and the design constraint he placed on it is almost old-fashioned: it has to be able to explain itself.
"The concept of being responsible and being able to truly understand decisions made by AI with an audit trail is extremely important in health care," he has said. The worry is specific. "As humans, we have the ability to explain the thought process behind our decisions, whereas many AI systems don't always have the documentation to explain why every decision was made." For a recommendation engine deciding what to watch next, that is a shrug. For a system shaping a person's recovery, it is a liability you cannot ship.
So the engine reasons out loud. As Anantha describes it, "among the many factors, it analyzes each patient's usage history with the product, compares that to the therapy progression of other patients like them, while also prioritizing what clinicians are assigning to their patients." Three signals - your own history, the wisdom of thousands of recoveries that look like yours, and the explicit judgment of your clinician - braided into a single next exercise. The clinician stays in the loop, never out of it.
That insistence is why the company can make a claim few in digital health can match: every one of those 100 million-plus exercises is data, and together they form what Constant Therapy calls the largest real-world brain rehabilitation database in existence. Anantha did not set out to build a dataset. He set out to help people, and the dataset is what helping people at scale leaves behind. The two now feed each other - more recoveries sharpen the engine, and a sharper engine produces more recoveries.
The path to a brain therapy company runs, improbably, through physics and the fastest silicon money could buy.
He started at the Indian Institute of Technology in Bombay, then crossed an ocean to Northwestern University, where he picked up a master's in physics and a PhD in electrical and computer engineering. Physics first, engineering second - a tell. He wanted to understand why things work before learning how to build them, and that order of operations shows up later in a man who refuses to let his AI act without a reason.
His first job was Lead Engineer at Motorola, building mobile software and hardware in the era when "mobile" still meant a phone that mostly made calls. He became Vice President of Engineering at a startup that Motorola then acquired, where he built software now used worldwide to manage wireless networks - the invisible plumbing that keeps signals flowing. Another venture, where he helped develop the world's fastest digital signal processor, was acquired by Apple. Two companies, two acquisitions, a fistful of patents. By most measures, a finished career.
He chose to start over in the hardest possible domain instead. Today he also teaches what he learned: a guest lecturer on entrepreneurship at Boston University's Questrom School of Business, an expert mentor at MassChallenge HealthTech and Insight Data Science, and a Charter Member of TiE Boston, the network that later honored him for entrepreneurial achievement.
At Constant Therapy Health, I think we can materially change the way healthcare works.
We wanted to build an AI system that would be accountable and explainable, that would respect patients' privacy, and that would make the most intelligent decisions with the information available.
The concept of being responsible and being able to truly understand decisions made by AI with an audit trail is extremely important in health care.
Which is why we built the NeuroPerformance Engine the way we did - to be accountable and explainable.
A clinic visit lasts an hour. A week has 168 of them. The gap between is where most brain recovery is won or lost, and historically it has been the part of medicine that technology ignored.
Anantha's bet is that the gap is exactly where software belongs. Put a clinically grounded, AI-personalized therapy program in someone's hand, and the rare specialist's expertise stops being something you visit and becomes something you carry. The work that earned Constant Therapy recognition from the American Heart Association and AARP was not a flashy demo. It was patient impact - the unglamorous, repeated, measurable business of people doing one more exercise than they would have done alone.
It is a worldview that fits the man. He spent decades optimizing systems that move information faster than humans can perceive. Now he applies the same discipline to a process that is the opposite of fast - the patient, halting, deeply human work of a brain rebuilding itself. The throughline is not speed. It is leverage: take something scarce and expensive, and make it abundant and personal. He did it with network software. He is trying to do it with recovery.