Scientific Computing & Spatial Intelligence
Enterprise experience meets cutting-edge research
- Astronomical Computing — Large-scale survey analysis, ML classification, interactive 3D visualization
- Machine Learning — PyTorch frameworks for ethics, relativity-inspired architectures, and fine-tuning
- DuckDB Extensions — Spatial, graph, astronomical, and space-filling curve extensions
- Scientific Tools — Paper search, reactive notebooks, AI-assisted UX for researchers
- Computer Vision — Open vision encoders, generative models, earth observation
| Repository | Description |
|---|---|
| astro-lab | Astronomical ML framework with advanced data analysis and interactive 3D visualization |
| astro-duck | DuckDB extension for astronomical calculations — coordinates, photometry, cosmology |
| earth2studio | Deep-learning framework for AI weather and climate workflows |
| illustris | IllustrisTNG public data release — cosmological simulation tools |
| marimo-flow | Interactive ML notebooks with reactive updates, AI assistance, and MLflow tracking |
| paper-search-mcp | MCP server for academic paper search across 11 platforms with SurrealDB knowledge graph and Docling document processing |
| Repository | Description |
|---|---|
| torch-cuda | PyTorch CUDA 12.8 project template — GPU-accelerated ML with modern packaging |
| torch-relativistic | PyTorch extension with neural network components inspired by relativistic physics |
| ethics-model | Modular PyTorch framework for ethical text analysis and manipulation detection |
| bidora | BiDoRA — Bi-Level Optimization for parameter-efficient LLM fine-tuning |
| OpenVision | [ICCV 2025] Fully-open family of advanced vision encoders for multimodal LLMs |
| Repository | Description |
|---|---|
| duckpgq | SQL/PGQ support and graph algorithms for DuckDB |
| lindel | Linearization/Delinearization — Z-Order, Hilbert and Morton curves |
| duckdb-spatial-3d | Geospatial functions, GEOMETRY type, and columnar spatial data |
Maintained by @bjoernbethge ·