Senior Scientific Software Engineer
Geospatial Modeling • Research Software Engineering • System Architecture
I build scalable scientific and geospatial software systems that support climate-resilient agriculture, high-throughput modeling workflows, and reproducible research. My experience spans NIH-funded computational modeling, large-scale data engineering at Climate LLC (Bayer Crop Science), and the development of open-source tools that strengthen the scientific software ecosystem.
- Distributed microservices (Python, FastAPI, async systems)
- Event-driven execution (AWS Lambda, Step Functions, SQS/SNS)
- High-throughput batch computation and chunk-based orchestration
- API design for research and production (REST, GraphQL)
- Raster workflows (GDAL, Rasterio)
- ROI clipping, zonal statistics, tiled computation
- Climate dataset integration and agricultural analytics
- Predictive diagnostics (WAIC, LOO, PSIS)
- Uncertainty quantification and posterior comparison
- Explainability pipelines (SHAP, partial dependence, calibration)
- Reproducible scientific workflows and auditability
- Containerized environments (Docker)
- CI/CD with GitHub Actions
- Observability (Loki, Promtail, Grafana)
- Self-hosted compute and storage (TrueNAS, Portainer)
- TerraFlow-Agro – Reproducible geospatial modeling workflow (submitted to JOSS)
- TrialFlow-Agro – Bayesian diagnostics toolkit for agricultural field-trial analytics
- ExplainFlow – Explainability and model auditing framework
- GeoVizFlow – Tools for reproducible geospatial visualization
- Computational Modeling Suite – Unified scientific workflow and container ecosystem
All projects emphasize reproducibility, transparency, and scientific rigor.
My work focuses on building open-source scientific and agricultural modeling systems that improve reproducibility, enhance model transparency, and support large-scale environmental and research decision-making across the United States.
Website: https://marupilla.dev
GitHub: https://github.com/gmarupilla
LinkedIn: https://www.linkedin.com/in/gmarupilla


