Zeaus leads a team of technical youth to build software and hardware to tackle a range of environmental challenges.
By 16, Zeaus had taught himself how to code and build AI models.
Seeing the rise
in environmental catastrophes due to climate change and the struggles of frontline
responders, he built a novel adaptive retrieval-augmented generation system to convert
natural language into reliable drone code.
This system, now referred to as
RescueAI, can enhance emergency responses to natural disasters such as floods, hurricanes
and earthquakes. The system performs 15-fold better than GPT-4-1106.
Zeaus was
recognised nationally as a Young Defence Scientist for his work and was personally
congratulated by the Minister of Defence. RescueAI is now being deployed for emergency
response missions. Click here to access news article.
In response to haze outbreaks and rising air pollution levels, and the high prices and
limited scope of commercial air purifiers, Zeaus and his team developed Filtrair, the first
solar-powered air filtration unit for outdoor use.
Filtrair can be assembled for
less than $50 and thus represents a solution anyone can use, anywhere. Zeaus was thus named
an Emergent Ventures Fellow (link) with a $20K grant to scale Filtrair.
He is also backed by the Singapore government.
Technology can be powerful, but
teaching youth how to build technology so they too may protect our Earth is even more
powerful. Zeaus has conducted workshops around Southeast Asia to teach youth how to build
their own Filtrairs.
Zeaus and his team is working on their biggest technical project to date, but very likely,
the most impactful. Global warming can no longer be solved with policy, there needs to be
effective ways to remove carbon dioxide from air.
Since 2024, we have been
exploring extracellular RuBisCO enzymes as a basis for a carbon sequestration unit. So far,
Zeaus managed to computationally optimise RuBisCO for 4-fold higher carbon capture than
typical plant enzymes.
Previously, we trialled RuBisCO in a hydrogel setting with
limited success, but are now working on using an aqueous environment of precursor solutions,
which has so far proven promising. Support our work!