👋 Hello, I'm Suriyah Dhinakaran

🚀 I’m a Geospatial Data Scientist working at the intersection of Earth Observation, AI, and Big Data, with a passion for enabling sustainable environmental development through open science and digital twin technologies.

🏢 Currently, I work as a Junior Researcher at the Institute for Earth Observation – Eurac Research, Bolzano, Italy 🇮🇹, where I specialize in climate data downscaling, Earth observation workflows, and high-performance environmental computing.

🧠 My work bridges climate modeling, machine learning, and reproducible research practices. I contribute to international projects like Horizon Europe – interTwin, support ESA-aligned workflows, and advocate for FAIR data principles in environmental modeling.

🎓 I earned my Master’s degree in Geoinformatics and Spatial Data Science under the supervision of Prof. Edzer Pebesma at the University of Münster, Germany 🇩🇪, where I focused on reproducible geospatial workflows and open science. During this time, I contributed to the Spatio-Temporal Modelling Lab, extending the open-access book Spatial Data Science with Applications in R by developing Python equivalents for broader accessibility.

🎯 Current Focus

• 🛠️ Scalable EO workflows with STAC + Zarr + openEO
• 🌡️ Climate downscaling using ML & ESRGAN
• 🔬 Digital twin applications for Earth system modeling
• 🤝 FAIR data, reproducibility, and open science

💼 I lead or contribute to the development of open-source tools such as:

• downScaleML – high-performance ML downscaling for climate data
  (main development happens in the interTwin EU GitLab)

• openeo-processes-dask – enabling Zarr-native processing and STAC integration
  (used in local, scalable EO pipelines)

• raster2stac – automated STAC metadata generation for EO rasters
  (developed within the internal GitLab of Eurac Research)

Most of my core development takes place on GitLab, and this GitHub space serves as a landing page for selected tools, experiments, and community-facing collaborations.

🧭 Professional Highlights

• 💡 Developed a two-stage ML downscaling method improving SEAS5 forecast resolution from ~30km to 1km
• 🛰️ Contributed to ESA’s EOPF Zarr service for Sentinel satellite data
• 🔄 Built raster2stac, streamlining metadata generation for FAIR EO data
• 🧪 Presented research at EGU, IEEE IGARSS, and won hackathons for EO-based ML solutions

📫 Get in touch

📊 GitHub Stats