Climate AI
Machine learning systems for weather, climate risk, and environmental intelligence.
AI research and engineering
AI systems for scientific discovery, climate resilience, and geospatial intelligence.
Editable.ai is a Pakistan-based AI research and engineering company building applied machine learning systems for scientific and social-impact domains.
Our current focus is climate AI, severe-weather modeling, geospatial data pipelines, and open scientific tools.
Climate AIsevere weather
Geospatial MLlarge-scale data
Open sciencebenchmarks
What we work on
Scientific AI with real-world context.
We build practical AI workflows for scientific datasets, climate-risk applications, and geospatial intelligence.
Geospatial Modeling
Large-scale pipelines for reanalysis, radar, forecast, and event-based datasets.
Open Scientific AI
Reusable models, benchmarks, and tools designed for research and public benefit.
Featured project
OpenSCS Foundation
Open foundation models for severe convective storm science.
OpenSCS aims to build reusable AI models and datasets for hail, tornado, and thunderstorm-wind hazards by integrating ERA5, ECMWF, MRMS radar products, and NOAA Storm Events reports.
Current stage
Proof of concept with preliminary models, hindcast visualizations, and spatial evaluation metrics.
Intended use
Scientific research, benchmarking, education, and climate-risk analysis. Not intended for operational warnings.
Training period
Prototype models shown on the OpenSCS page use a 2000–2015 training period.
Test period
Current proof-of-concept figures summarize evaluation over 2016–2025.
Proof of concept
Early severe-weather model outputs.
These figures show preliminary severe-weather model pipelines, including label generation, balanced sampling, U-Net-based model training, inference, hindcast visualization, and spatial evaluation.
Click to enlargeHail Hazard Model
Regression model output for severe hail hazard visualization.
Click to enlargeTornado Probability Model
Rare-event probability model output for tornado hazard research.
Click to enlargeThunderstorm Wind Hazard Model
Regression model output for damaging thunderstorm-wind hazard visualization.
Project Lead
Syed Sabbih Haider Shah
AI Research Lead / Climate AI Researcher. Sabbih works on climate AI, severe-weather modeling, geospatial ML pipelines, and scientific deep learning. His work includes severe-convective-storm models for hail, tornado, and thunderstorm wind using ERA5, MRMS, and storm-report-based labels.
Open-science focus
Editable.ai is preparing OpenSCS as an open scientific AI effort focused on reproducible workflows, clear evaluation, model cards, data cards, and responsible-use documentation.
Contact
Contact Editable.ai.
For project questions, collaboration, or review, contact sabbih@editable.ai.