Vision
The same computational thinking that drives my research — applied to ecology, traditional knowledge, and the living world.
A smart greenhouse as a first tangible step toward computational ecology. The long-term question: can AIVC concepts generalize from cells to ecosystems?
Scoping the hardware stack: DHT22 for temperature/humidity, MH-Z19 for CO₂, soil moisture array. The goal is to treat the greenhouse as a biological system with AnnData as the data format — time series become observations, sensor channels become features.
● In progressComing soon — configuring edge compute, writing the sensor polling daemon, and designing the AnnData schema for time-series ecological data.
The long-term goal: applying AIVC-style universal representations to ecological time-series data. Can we learn meaningful "cell states" for plants in a shared environment?
Bridging traditional herbal knowledge and modern perturbation biology — from field collection to Virtual Cell prediction.
Medicinal herb specimens collected and catalogued. Traditional use cases documented as biological hypotheses.
Liquid chromatography-mass spectrometry identifies active compounds and metabolite profiles from raw plant material.
Compounds mapped to LINCS L1000 gene expression signatures — connecting chemical structure to transcriptomic response.
AIVC framework predicts cellular response to herbal compounds — bridging ethnobotany and computational systems biology.
Can traditional herbal knowledge — accumulated over centuries of empirical observation — be formalized as perturbation biology hypotheses and validated computationally before wet lab synthesis?
Initial target: oral cancer biology. Field-collected medicinal herbs with known anti-inflammatory or anti-proliferative traditional uses, evaluated against oral cancer cell line transcriptomes via AIVC.
Exploratory research phase. Pipeline design complete. Awaiting LC-MS access and initial compound panel.