Research
My research examines how geospatial data and applied AI can support decisions in data-constrained, rapidly changing environmental contexts.
The work combines methodological development with reproducible implementation, domain knowledge and collaboration.
Rather than treating models as isolated outputs, I focus on the full research chain: data, assumptions, evaluation, transferability, evidence and practical interpretation.
Research themes
Optical, SAR, elevation, time-series and multisource geospatial data for land-use and land-cover mapping, agriculture, forest and mangrove monitoring, environmental change analysis and spatial prediction.
Methods: CNNs, UNet/UNet++, semantic segmentation, ensemble learning, supervised and unsupervised learning, time-series analysis and transferability evaluation.
Applied AI and geospatial research supporting climate-resilient agriculture, blue carbon, ecosystem monitoring, hydrology, flood and drought analysis, and MRV-oriented environmental workflows.
Research and prototyping around retrieval-augmented generation, source traceability, citation and claim validation, numerical consistency, deterministic verification and local or remote language-model integration.
Reproducible Python workflows, APIs, automated tests, environment management, Docker packaging and technical documentation designed to make research outputs easier to evaluate, reuse and extend.
Current research
Current work includes geospatial and satellite-data workflows, AI models, multidisciplinary technical delivery, project documentation, proposal development, partner engagement and research communication.
Public outcomes will be added as they are approved.
Evidence