A Cochlear Implant Gave His Mom Her Hearing Back. He Built Ketryx So the Next Breakthrough Reaches Patients Faster.

Regulated medical product development runs on proper documentation. A single medical device can take tens of thousands of pages to explain, and every software update multiplies the burden. "If you want to ship this product every week because you have a software update, and it takes 10,000 pages to do it, you would need, for one product in one region of the world, to create 500,000 pages of evidence," says Erez Kaminski, CEO and founder of Ketryx. "That's just not feasible to do." Teams must choose between speed or compliance. For patients waiting, that tradeoff has real consequences.

Ketryx began with a question its founders spent two years answering: why does compliance take so long? As an MIT grad student in Cambridge in 2019, Kaminski read tens of thousands of pages of regulation, mapping how it all connected. After a year of research and then another year of building, he produced the first version of the platform. The insight was specific: if AI agents could interoperate with the development tools regulated teams already use, analyzing work as it happens and capturing compliance context automatically, then regulated development could move at the same speed of regular software development. Reaching that goal meant building for scale and for critical use cases, where an error is not a bug but a risk to a patient.

That raised the bar on infrastructure. "One of the reasons to use AWS is because the way they run their quality system actually is much more appropriate for highly regulated companies compared to most cloud providers," says Kaminski. Ketryx needed more than a cloud provider. It needed a collaborator on a hard problem: how to build AI-powered compliance automation that could operate at scale in environments where the cost of failure is measured in patient outcomes. "We had someone to collaborate with on how do we build something that could be used at such a big scale and for such critical use cases," Kaminski says, "and I think all of our partners feel the same way."

Ketryx connects to the development tools regulated teams already use and deploys AI agents that analyze the work as it happens. Instead of asking engineers to stop building and start documenting, the platform reads from the tools where the work lives and generates the records that regulators require. The output is audit-ready evidence, produced continuously rather than reconstructed at the end of a release cycle.

The proof came from a customer. "One of the moments that made me understand that we're really onto something is the first time a team connected us to many different IT systems and generated all their compliance documentations and metrics," says Kaminski. "Then they sat and reviewed it and told us, this is pretty amazing. It saved us 90% of the documentation time."

That result matters because of what it replaces. What once consumed most of a team's time now happens in the background, and the output holds up where it counts. For a regulated company, passing an audit is not a nice-to-have. It’s the difference between shipping and stalling.

Ketryx runs on AWS for a reason that goes deeper than compute. Building AI agents for regulated industries means the infrastructure itself must meet the same bar as the customers it serves. Kaminski and his team evaluated cloud providers against the compliance and quality standards their own customers are held to. AWS's approach to quality systems and controls fit the needs of highly regulated companies in a way other providers didn't.

That foundation lets Ketryx make a promise it can keep: the AI platform itself operates within the kind of controlled environment its customers must demonstrate. By meeting teams inside the tools they already trust, Ketryx removes the tradeoff between speed and compliance. Every change is captured as it happens, so the evidence trail is continuous rather than reconstructed at the end.

The scale question was central from the start. Ketryx set out to serve products that require hundreds of thousands of pages of evidence across regions and releases, which is a volume no manual process can sustain. AI agents that automatically capture compliance context are the only way to make that math work.

What the customer gets is straightforward. Engineers spend their time building the product instead of documenting it. Quality and regulatory teams get evidence that’s audit-ready by default. And the organization gets to ship regulated software on the cadence modern development demands, rather than the cadence paperwork allows. The 90% reduction one customer reported isn’t a marketing figure. It’s time returned to the people whose work brings devices and therapies to patients faster.

The measure of Ketryx isn’t the platform. It’s the patients on the other side of it. "The idea that we're already today reaching millions of patients is beyond my wildest dreams," says Kaminski.

For Kaminski, the mission is personal. "My mother is deaf. She went from a person who was quite profoundly hearing impaired to a person who can listen and talk, because she went through cochlear implant surgery. Through that surgery, we experienced the amazing miracle that is biomedical innovation." That experience is the reason the company exists and the reason speed is never abstract.

What’s next is more of what’s already working: more teams generating audit-ready evidence automatically, more regulated products reaching patients at the speed of modern software, and a platform built to scale with the industry's most critical use cases. "The evidence is A-grade, it's passing our internal audit, and we just want more of it," one customer said. For Ketryx, that’s the whole point.

Kaminski went from a napkin sketch in a Cambridge apartment to an AI platform reaching millions of patients. He built it on AWS because regulated products need infrastructure that earns trust at scale.

If you're a founder with an idea that could change how your industry works, AWS AI can help you get from concept to working product faster. Explore the AI for Startups hub to access foundation models through Amazon Bedrock, accelerate development with Kiro, and apply for up to $200K in AWS Activate Credits to start building today.