What pivoting Illume Bio from AMR diagnostics to drug discovery taught us
By Kushal Chandak and Sambhav Baid, co-founders, Illume Bio
We started Illume Bio with a problem that genuinely kept us up at night.
In most hospitals, a doctor knows a patient has a serious infection long before they know which antibiotic will work. Culture-based susceptibility testing takes two to three days. In sepsis, that is not a delay, that is a decision made without information. So doctors do the only rational thing available: reach for broad-spectrum antibiotics and hope. Repeated millions of times a year, that habit is one of the biggest engines of antimicrobial resistance in the world.
Our idea was to close that gap. Instead of asking whether a bacterium carries a resistance gene, which is what PCR does, we wanted to ask a sharper question: is the resistance enzyme actively destroying the antibiotic right now? We designed a synthetic probe that stitched an antibiotic to a DNA barcode through a self-immolative linker. If the resistance enzyme was active, it would cleave the antibiotic, the linker would fall apart, and the released DNA would amplify into a fluorescent signal. Enzyme activity in, digital readout out. Under an hour, no culture needed.
We loved it. It was elegant, it was ours, and on paper it worked. We do not work on it anymore. Here is why.
The product was solvable. The path to a patient was not.
Every hard question we asked ourselves in the early months was technical. Can we make the linker chemistry work? Will the signal amplify enough? Can we cut background in serum? Those were real questions, and they were answerable. Chemistry is stubborn, but it negotiates.
The questions we did not ask early enough decided our fate. What does the regulatory pathway for a novel diagnostic actually look like, in years and in capital? Who signs the purchase order in a hospital, and how long is that cycle? What does it take to displace a machine already sitting in the lab, already validated, already inside someone’s workflow? Who pays for the test?
None of those had good answers for a two-person team with no wet lab. The science was not wrong. The shape of the market simply did not fit the shape of us.
That is the lesson we would put on the wall: feasibility is not viability. You can prove a thing is possible and still have no business. Founders like us are trained to solve, so we spent months answering “can we build it?” when the real question was “can we build it, sell it, and survive the wait?”
Talk to people in the field before you fall in love
The fastest and cheapest experiments we ever ran were conversations. Clinicians, microbiologists, people who had taken a diagnostic to market and lived through it. Twenty minutes with the right person taught us more than three weeks of literature, and a few of those calls were uncomfortable in exactly the way useful feedback is. Validate the need, then the scope, then the science. We did it backwards and paid for it in months.
A pivot is not a reset. It is a transfer.
The most freeing realisation was that pivoting did not delete a year of work. Almost everything carried over: the molecular design thinking, the DNA engineering, the modelling instincts, the discipline of benchmarking every claim against published data. All of it walked with us into drug discovery, where we now build computational pipelines for aptamer design. The diagnostics work is not a scar, it is compound interest. We are faster at the new problem precisely because we were slow at the old one.
There will be more pivots. Get good at them.
We used to think of a pivot as an emergency exit. We do not see it that way now. A pivot is what learning looks like when you are honest enough to act on it. If your view of the world has not changed after a year of building, you probably were not paying attention. So the goal is not to avoid pivots, it is to build the muscle for them.
What we would tell a founder starting today
- Fall in love with the problem, but interrogate the market before you commit years to the solution.
- Ask “is this solvable?” and immediately after, “is it solvable by us, at our size, in our timeline?”
- Treat every conversation with a practitioner as an experiment. Highest return per hour you will get.
- When you pivot, carry your assets with you. Skills, tools, relationships and judgment are all portable.
We are still building. The problem is different, the ambition is not. And we are far more useful now than we were when we were certain