An open-source visual analytics workbench to explore, search, and validate geospatial embeddings and atmospheric foundation model latent spaces.
Comprehensive evaluation of WRF model rainfall prediction skill during Super Cyclone Kyarr (2019) using GPM & TRMM datasets. Includes day-wise and threshold-wise metrics, visualizations, and automated report generation.
Deep learning statistical downscaling of climate model outputs using CNN super-resolution on ERA5 data
Summaries of research in climate, geoscience, and climate-adjacent applications of AI
Dynamic Cyclone Early Warning System using LSTM and IBTrACS dataset
Software Developer & Nexa Founder — Building computer vision and geospatial AI for climate adaptation and social impact.
StormFormer: physics-informed hybrid Transformer-BiLSTM for 24h typhoon track prediction on IBTrACS
Autonomous Spatial Heat Stress Intelligence & Operator Command Center - FortyGuard Hackathon 26
Streamlit ML app predicting cyclone intensity from SST, wind shear, and pressure using gradient boosting. Storm-track visualisation on Folium interactive maps.
Adaptation of NVIDIA StormCast for India by replacing the HRRR data pipeline with IMD GFS and WRF preprocessing, synchronization, training, inference, and visualization workflows.
Agente de IA focado em previsão de desastres e contenção de danos causados pelas discrepâncias temporais no estado da Bahia
Climate-aware crop recommendation system with a reproducibility-ready public artifact, evaluation outputs, and a documented reconstruction pathway.
AI-powered flood risk prediction for Nyando Basin, Kenya. XGBoost + SHAP + FastAPI + React. AUC-ROC 0.9717. 100% open data — Kenya DPA 2019 compliant.
Self-supervised deep learning for rainfall super-resolution: downscaling CHIRPS precipitation from 0.25° to 0.05° over Kerala using an SRResNet-style GAN with PatchGAN discrimination.
Enterprise-style hybrid AI assistant combining structured climate analytics, RAG, hybrid retrieval, reranking, and conversational memory using local open-source LLMs.
Open-source environmental intelligence and AI research for extreme heat, PM2.5, air quality, wildfire smoke, urban climate, forecasting, uncertainty, and resilient homes.
An autonomous multimodal AI framework for disaster prediction, satellite flood segmentation, physics-informed validation, ensemble fusion, and adaptive monitoring across earthquake, cyclone, wildfire, and flood events.
NVIDIA Earth-2 climate AI workspace: a structured set of exercises, notebooks, and templates for earth2studio.
Spatial-temporal machine learning and geostatistical Kriging interpolation for high-resolution regional temperature forecasting and environmental monitoring.
Private research repository.
To associate your repository with the climate-ai topic, visit your repo's landing page and select "manage topics."