Sebastian Galkin

[Sebastian Galkin @paraseba]

The US National Weather Service is adopting Icechunk!

[Earthmover @EarthmoverHQ]

@NOAA's NWS is adopting Icechunk for CIRRUS, the next-gen system replacing AWIPS.

🧊 ACID transactions + cloud-native I/O at PB scale
🚀 100x faster than existing NetCDF workflows
🏛️ Serving FAA, FEMA, NASA, USDA, DOE + every state

https://t.co/0Tc8kvDgJI

[Sebastian Galkin @paraseba]

Very proud of the work we are doing. Free cloud datasets for science, SLAs for businesses, best consistency and performance we know of. See also our recent ERA5 dataset https://t.co/hwLyIESpO0

[Earthmover @EarthmoverHQ]

🛰️ Old file format? No problem! We made the full GOES-16 satellite archive queryable as a single Zarr store — without copying a single byte 🧊 Find out how: https://t.co/tQ37HkMApA

[Earthmover @EarthmoverHQ]

🛰️ Old file format? No problem! We made the full GOES-16 satellite archive queryable as a single Zarr store — without copying a single byte 🧊 Find out how: https://t.co/tQ37HkMApA

[Earthmover @EarthmoverHQ]

Ship fast, break nothing. 🚀

We built a test that runs Icechunk v1 AND v2 in a single Python process, randomly interleaves their calls, and auto-shrinks failures — using a tool called third-wheel 🛞

Result: zero v1/v2 compatibility issues reported. ✅

Full write up link👇

[Earthmover @EarthmoverHQ]

🌊 Arraylake now speaks MCP.

Connect Claude, ChatGPT, or Cursor to your array data to browse repos, query Flux, render maps, all in natural language. Agents empowerment, no notebook required. ⚡

Full announcement & links to the docs👇

[Earthmover @EarthmoverHQ]

🧊 Icechunk V2: made more reliable for the scale V1 users wanted with help from great OSS tools.

🛤️ shuttle (@awscloud) systematically explores async task orderings to find race conditions

☢️ toxiproxy (@Shopify) poisons network connections to harden retry logic

Links 👇

[Earthmover @EarthmoverHQ]

From GeoTIFF chaos → cloud-native climate risk: Eoliann built AIRIS to predict floods, fires & droughts with precision across time horizons to 2050.

Powered by Earthmover, they turned TBs of satellite data into fast, decision-ready insights: https://t.co/FwOvZpETYZ

[Earthmover @EarthmoverHQ]

Icechunk V2 is coming soon- so to check compatibility with V1 pre-release, @xarray_dev Ian Hunt-Isaak built third-wheel, a tool to 𝐢𝐧𝐬𝐭𝐚𝐥𝐥 𝐦𝐮𝐥𝐭𝐢𝐩𝐥𝐞 𝐯𝐞𝐫𝐬𝐢𝐨𝐧𝐬 𝐨𝐟 𝐭𝐡𝐞 𝐬𝐚𝐦𝐞 𝐩𝐚𝐜𝐤𝐚𝐠𝐞 𝐬𝐢𝐝𝐞 𝐛𝐲 𝐬𝐢𝐝𝐞: witness the magic.

[Earthmover @EarthmoverHQ]

Announcing the Earthmover Data Marketplace: cloud native access to high-velocity weather, climate, & E/O datasets. Get access to ARCO ready data cubes from trusted providers, including forecasts from @ECMWF @NOAA- and exciting new AI-based forecasts!

[Earthmover @EarthmoverHQ]

🧊 Icechunk 2 is now available in alpha: Move/rename nodes without rewriting, shift chunks without consuming I/O, write metadata at the repo level, host repos on any HTTP server, and so many performance improvements. Read @rabernat and @paraseba's post for the full story! 🎁

[Earthmover @EarthmoverHQ]

In case you missed it: Tomorrow @cherian_deepak will give a live demo that will take you from Zarr/Icechunk to Mapbox/MapLibre with a click of a button!  Build faster, smoother maps with no pre-processing using our latest feature: Tiles. https://t.co/57nBKgIiJx

[Earthmover @EarthmoverHQ]

🧊 Icechunk 1.0 is production-ready!

Deployed by weather & climate data teams for rapid operational access to data on cloud

⚡️ 100x faster reads vs. NetCDF
🌎 petabyte scale

Big day for #WeatherData teams!

👉 https://t.co/dAFQYOYSoV

#climateTech #OpenSource #meteorology

[CNG @cloudnativegeo]

Lindsey Nield & @cherian_deepak from @Earthmover's #CNG2025 talk is live on YouTube. Learn about Zarr, Icechunk & Xarray for cloud-native data cube analysis.
Watch now: https://t.co/PVxkqGpAAi

[Sebastian Galkin @paraseba]

@milenaarmada @sistemafierro Felicitaciones a todos! Fenomenal!!! Qué rápido pasa!

[Sebastian Galkin @paraseba]

After months of Rust, I wrote some Python this weekend. I immediately got burned by global mutable state

[Sebastian Galkin @paraseba]

Seriously blown away by this talk - likely the best I've ever seen on open data. @cherian_deepak gives a captivating recount of the historic context, clearly shows where we are falling short, and crucially, what we can do about it.

Don't miss it! https://t.co/60dOHlZpte

[Deepak Cherian @cherian_deepak]

I had a boatload of fun preparing and delivering this talk! Watch the recording here: https://t.co/Xa64nipAkD Please reach out if you have comments or other hot tips on this topic!

[Sebastian Galkin @paraseba]

Short blog post on the scalability of the multi-dimensional array store I've been working on for the last few months.

https://t.co/DFdrCZNwam

[Sebastian Galkin @paraseba]

I've worked on Icechunk almost exclusively for the last six months. I'm very proud of the result; you should check it out.

[Earthmover @EarthmoverHQ]

1/ 🚀 Solving #NASA ’s cloud data dilemma: Icechunk unlocks 100x faster access to archival data formats

We're thrilled to publish results from our pilot project with @NASA + @developmentseed to enable high-performance cloud-native access for NASA’s 100s of PBs of EO data.

[Sebastian Galkin @paraseba]

RT @rabernat: Most developers today wouldn't dream of not using version control for their code... However, the same principles can be appli…

[Pangeo @pangeo_data]

Join us tomorrow Oct 23 at 4 PM EDT for a Pangeo showcase on Icechunk: An Open-Source Transactional Storage Engine for Zarr by @rabernat! https://t.co/lsfzcXYaqn