AidBio did not begin with AI.
It began with a biological question: how do many small signals inside a living cell add up to something you can actually see or measure — a trait, a stress response, a yield? Two decades later, that's still the question underneath AidBio — now answered with bioinformatics, machine learning, scientific software, environmental data and AI-assisted research instead of a single lab bench.
The tools changed. The scientific method did not.
Mechanism first
Before modeling anything, ask the same questions a bench scientist asks: what are the variables, what interacts with what, what did we actually measure versus infer, and what could be misleading us? That habit came straight out of molecular genetics and hands-on plant biology.
Represent complexity honestly
Once a system has too many moving parts to hold in your head, turn it into structured data instead of guessing — gene activity tables, interaction networks, measured traits, weather records over time.
Make the evidence reusable
A one-off spreadsheet analysis only answers one question, once. Software that captures the same reasoning can answer the next hundred questions like it — for a client, not just for one paper.
From crop biotechnology to scientific AI
The story starts in applied agricultural biotech at Alellyx, deepens through years of academic research on how plants sense sugar, hormones and stress, expands into mapping how proteins interact inside crop cells, and then moves increasingly into computation — graph-based machine learning, climate analytics and AI-assisted research software.
The important point is continuity, not a career change: the move into computing wasn't an escape from biology. It was what biology demanded once the questions got too complex to answer by hand.
Crop genetics in industry: finding the genes behind useful traits and validating them in the greenhouse and field.
PhD and postdoctoral research on how a single regulatory protein (bZIP63) lets a plant's energy and stress signals talk to its internal clock — plus the RNA and protein-interaction data behind that finding.
Plant immunity, protein-interaction networks, rice stress biology, and turning plant images into measurable data at scale.
Scientific software, bioinformatics, graph-based machine learning, agricultural climate analytics, and AI workflows that keep the underlying evidence traceable.
“Can molecular, biological, agronomic and environmental complexity be transformed into reproducible computational evidence that researchers can actually use to make decisions?”
This is the question AidBio is structured around.