K-NetZero Compass
A deployable decision-support demo for net-zero planning. Built to be transparent, auditable, and extendable (future LCA/TEA hooks and uncertainty-aware comparisons). (Free tier may sleep/wake.)
I build law-constrained, deployment-ready models for wastewater treatment plants: forecasting, anomaly detection, and climate-stress resilience — bridging process engineering, microbial systems, and Scientific Machine Learning.
My work sits between hard-core process engineering and reliable AI. I build models that respect physical/biological constraints, remain robust under plant realities (daily sampling, sensor drift, wet-weather shocks), and translate predictions into decision-relevant signals.
Operational SciML for resilient water infrastructure: forecasting + anomaly detection + resilience metrics, grounded in process memory, kinetics, and climate-regime shifts. In parallel, I contribute to sustainable carbon management (CCUS/ICCU).
Operational forecasting, anomaly detection, and decision-ready monitoring across multi-contaminant signals and operational states.
Regime-aware evaluation for wet-weather shocks, heatwaves, drought/low-flow, and recovery dynamics with interpretable metrics.
Integrated capture–conversion thinking, audit-ready TEA/LCA hooks, and experimental sorbent work aimed at realistic conditions.
A deployable decision-support demo for net-zero planning. Built to be transparent, auditable, and extendable (future LCA/TEA hooks and uncertainty-aware comparisons). (Free tier may sleep/wake.)
A set of small, deployable prototypes: anomaly monitoring dashboards, resilience scoring (4R/CRPS), and carbon-accounting workflows designed for reproducibility and real adoption.
Multi-contaminant forecasting and anomaly detection for real plant conditions: daily sampling constraints, wet-weather shocks, regime shifts, and sensor drift. Outputs are decision-ready (risk-of-violation, early warnings, recovery metrics).
A framework to connect operational trajectories (DO/ORP, loading, salinity, temperature, wet-weather memory) to microbial pathway indicators (nitrification/denitrification/Anammox proxies), designed to integrate metagenomic/16S features as “biological sensors.”
Predictive monitoring for AD stability (VFA/alkalinity/pH dynamics), early warnings for upsets, and controllable levers for energy-positive operation. Built to be compatible with lab-scale reactor datasets and biotech workflows.
Published review on sustainable CO₂ capture, conversion, and integrated strategies. Experimental work in preparation on nanoparticle-augmented diatom biosilica sorbents: humidity/temperature robustness and recyclability across cycles.
A deployed prototype for net-zero decision support. Designed as a foundation for uncertainty-aware comparisons and audit-ready reporting.
A decision layer that converts predictions into operational resilience: time-in-safe, headroom, recovery, and actionable intervention signals—built to remain interpretable under stress regimes.
From Emissions to Assets: Sustainable Technologies for CO₂ Capture, Conversion, and Integrated Strategies
International Journal of Molecular Sciences (2026). DOI: 10.3390/ijms27020847
Nanoparticle-Augmented Diatom Biosilica as a Sustainable Sorbent for CO₂ Capture (in preparation)
Baseline performance, humidity/temperature effects, concentration dependence, and recyclability across cycles.
Email: [email protected]
ORCID: 0009-0003-0412-3112
Note: For supervisors, a minimal page wins: affiliation, 2–4 strongest projects, and DOI-backed publications.