IsobarNWP archive

About us

The data existed. Reading it did not.

Isobar started when a hydrology group needed forty years of hourly precipitation at two hundred points and was quoted six months.

2016
founded
222
people
20
countries
100%
engineer-owned

Origin

Chunking is the whole problem

Model output is written for the forecaster: one field, one time step, one file. Almost every scientific question wants the transpose — one point, every time step. Storing both layouts costs a lot of disk and saves a lot of years, so we did that.

  • Both chunkings kept for every dataset that justifies it
  • Original GRIB retained byte-identical
  • Cycles kept after the originating centre deletes them
2018

One model, one university

ERA5 rechunked for hydrology.

2020

Operational ingest

Cycles online within the hour.

2022

Ensembles

All members, not just the control.

2024

Twelve-minute lag

Streaming ingest replaces batch.

2026

18.2 PB

Nine centres, 9.1 TB a day.

How we work

Boring on purpose

We would rather run software that is three years old and understood than something new that pages us at 04:00.

Own the path

We buy transit from multiple carriers and peer directly wherever it is possible, so a single bad hop is never a single point of failure.

Write it down

Every incident gets a public write-up. Every design decision gets an RFC. Institutional memory beats heroics.

Some of the people behind it

A quarter of the company has been here since the first rack.

G. MarchettiDeveloper relations
Y. TanakaProduct
D. HaugenPlatform
H. FerreiraDesign
A. VasquezPlatform
B. SørensenSecurity
U. BergströmDesign
S. BergerProduct

We are hiring

Remote-first across EU time zones, with a small office network for people who prefer one.