Add GeoTessera getting-started notes from Smayan

README.md

@@ -76,6 +76,7 @@ TESSERA is Cambridge's geospatial foundation-model project (Anil Madhavapeddy's
 #### Internal Docs & Notes
 
 1. [channel-files/TESSERA-understanding.pdf](channel-files/TESSERA-understanding.pdf) (Uttkarsh).
+2. [geotessera/README.md](geotessera/README.md) (Smayan Agarwal): learning resource documenting the process of getting started with GeoTessera.
 
 #### Project Management
 

geotessera/README.md

@@ -0,0 +1,61 @@
+# Getting Started with GeoTessera
+
+*A learning resource by Smayan Agarwal, documenting the process of learning GeoTessera. Ported from a [Slack post](https://fplaunchpad.slack.com/archives/C0B3GBWL97U/p1786178047916439) shared in `#fpl-esg`. Might be of help to others getting started as well.*
+
+## Best introduction
+
+[ESA: "Tessera AI model offers accessible way to view Earth"](https://www.esa.int/Applications/Observing_the_Earth/Copernicus/Tessera_AI_model_offers_accessible_way_to_view_Earth)
+
+IMO the best resource to introduce yourself to the idea of what GeoTessera even is.
+
+## Useful resources
+
+[Quick Start Guide](https://geotessera.readthedocs.io/en/latest/quickstart.html)
+
+- [Tutorials](https://geotessera.readthedocs.io/en/latest/tutorials.html) — covers downloading/analyzing embeddings in Python, and sampling embeddings at specific point locations.
+- [CLI Reference](https://geotessera.readthedocs.io/en/latest/cli_reference.html) — for the download/visualize/serve command-line workflow instead of the Python API.
+- [Architecture Guide](https://geotessera.readthedocs.io/en/stable/architecture.html) — once you're past basics, explains how the embeddings/foundation model are structured.
+
+Anil Madhavapeddy's own [announcement post](https://anil.recoil.org/notes/geotessera-python) gives good context/motivation for why the library exists.
+
+## Running the interactive notebook demo
+
+The announcement post also has an entire Jupyter notebook you can run locally. Mixed reviews — it's pretty cool, but generating layers for even a small area takes a LONG time. An area triple the size of Cambridgeshire took 15 minutes. Might be nice as a small demonstration thing though.
+
+Code to get the Jupyter notebook up and running:
+
+```bash
+git clone https://github.com/ucam-eo/tessera-interactive-map
+cd tessera-interactive-map
+uv venv
+source .venv/bin/activate
+uv pip install -r requirements.txt
+code app.ipynb
+```
+
+## Other demos and resources
+
+[tee.cl.cam.ac.uk/viewport_selector.html](https://tee.cl.cam.ac.uk/viewport_selector.html)
+
+A small demonstration of what GeoTessera is. Not very deep though. Low cost, low reward.
+
+[geotessera.org/docs/#tutorials](https://geotessera.org/docs/#tutorials)
+
+A listing of a selection of the resources that GeoTessera has.
+
+[isaac.earth/geospatial-skills](https://isaac.earth/geospatial-skills/)
+
+A library of Claude skills that will be useful.
+
+[Tessera announcement paper (arXiv:2506.20380)](https://arxiv.org/abs/2506.20380)
+
+## Cool applications of GeoTessera
+
+Two cool applications found:
+
+- <https://arxiv.org/pdf/2604.09818>
+- <https://arxiv.org/pdf/2311.16187>
+
+## Gotchas
+
+As of the time of downloading and trying to use things, downloading everything took a long time because of rate throttling on their end.