DiagnoBac’s

I never set out to build a diagnostics company.

My original vision was much broader. I started a company called BioForge because I wanted to build AI systems that could help us understand biology and keep making discoveries long after any individual scientist could. My belief was simple: I don't just want to make discoveries in my lifetime—I want to build a machine that keeps making discoveries long after mine.

As I worked on that vision, I realized I was trying to solve everything at once. Biology is simply too vast. If I wanted to make progress, I had to start with one fundamental problem.

Around that time, I read Brain on Fire. It reminded me of something I'd grown up seeing but had almost stopped noticing. My father was a government medical officer, and I had watched patients travel long distances just to access basic diagnostics. Later, while working on environmental sensing, I saw activists unable to prove river pollution before the damage was done. I saw the same thing again in food safety. Different industries, but the same underlying problem: we couldn't sense what mattered, when it mattered.

That sent me back to first principles.

The more I thought about biology, the more I realized that everything begins with sensing. Before a cell can respond, adapt, or make a decision, it first has to recognize what's around it. In many ways, the principles of sensing even predate life itself. Long before the first cell evolved, chemical systems were already responding to their environment. Life didn't invent sensing—it inherited and refined it. Every biological decision begins with molecular recognition.

Then something clicked.

The real bottleneck wasn't diagnostics. It wasn't biosensors.

It was that we still can't design molecular recognition systems.

We still discover them through massive screening campaigns, testing millions of candidates in the hope that one works. That isn't engineering. It's search.

That's when the company changed.

Instead of building another biosensor, I decided to build the missing design layer.

The question became:

Given a desired biological function, can we design the molecule that achieves it?

That became DiagnoBac's.

Today we're building AI models that learn the relationship between function, structure, and sequence, enabling intent-based molecular design. We started with molecular recognition because it sits at the foundation of biology, and biosensing is the clearest place to prove that this approach works.

Our first model, NucleoFold, designs nucleic acid capture molecules for biosensors.

Biosensors are our first wedge.

The real ambition is much larger.

We're building the design infrastructure that makes biology itself engineerable.