ChoLab

Cholera is waterborne, which means it spreads at the community level. Once water is contaminated, everyone who drinks from it is at risk. The existing detection methods are all reactive: by the time you get test results from a lab 5 hours away by car, the outbreak has already spread. There was no way to detect contamination proactively, before people got sick.

I led the product side of a 3-person team (designer-engineer, developer, myself as product lead) over 6 months. We combined LAMP isothermal amplification with a low-power heater, fluorescence-based dye, and a smartphone camera. The user adds a water sample to a test tube containing freeze-dried reagents, waits 20 minutes, and gets a result on their phone: safe or contaminated.

Freeze-drying was the cost breakthrough. Traditional lab methods require opened reagent supplies that expire and go to waste. Pre-dosed, single-use tubes eliminated that entirely, cutting the cost from $1,000 per test to under $10. It also removed the need for cold chain transport, which had made field deployment impractical.

We ran 50+ stakeholder interviews with WHO, Gates Foundation, UNDP, Harvard School of Public Health, and medical staff across India and Kenya. Field deployment started in West Bengal, where we tested with communities that had no prior access to water quality data.

Before building anything, I mapped the full system. 40+ factors from stakeholder interviews, literature, and field research, traced from root causes to downstream effects. The map narrowed to 11 intervention points where action could cascade through the network.

97% detection accuracy. Results in under 20 minutes. Over 85% of users with low literacy were able to complete the test independently. Field-deployed in West Bengal with real community use, tested across India and Kenya.

ChoLab won the iF Design Award Gold and was selected as a Global Top 100 Prototype by Dubai Future Foundation.

The biggest insight was understanding why existing systems were reactive in the first place. It wasn't a technology problem. It was a cost and accessibility problem. When a single test costs $1,000 and requires a lab, of course you only test after people are already sick. Changing the form factor changed the entire logic of when and why you test. That principle, that the shape of the solution determines when it gets used, has influenced how I think about AI deployment too.