Hacker News: lowsenberg

New comment by lowsenberg in "DeepInfer: High-performance inference framework for DeepSeek models"

lowsenberg — Tue, 20 Oct 2026 12:15:14 +0000

I'm hoping that hosting providers such as Hetzner will soon offer inference at reasonable prices and hosted in the EU with this kind of framework.

New comment by lowsenberg in "Ask HN: Who wants to be hired? (June 2024)"

lowsenberg — Tue, 04 Jun 2024 07:11:08 +0000

Location: Hanover, Germany

Remote: Yes

Willing to relocate: No

Technologies: Python, C/C++, databases, Django, HPC (MPI, threads), web technologies, gRPC, Docker, comp sci (numpy/scipy/HDF5/FFTW/....)

Résumé/CV: upon request

Email: hn@jomx.net

I am a theoretical Physicist (PhD, semiconductor research) and software developer. I have broad experience in scientific programming, web and application development with Python and the usual front-end stuff, and I've worked a lot with data storage and processing. At the moment I work for a large German car maker in the R&D department, but I'd like to do something more meaningful. My ideal job combines math/natural sciences and professional software development, with enough flexibility to spend time with my family.

New comment by lowsenberg in "Show HN: Cash: Tool for running shell commands on a large number of HPC nodes"

lowsenberg — Thu, 15 Sep 2022 05:51:45 +0000

I'm administrating a large HPC infrastructure at may day-to-day work and often need to check something on many or all of the nodes. Compare directories, files, system settings, or the like. As we have a total of around ~7000 nodes at different geographical sites, all other tools were unsatisfyingly slow when I want to run a command on all of them.

cash (with warm caches) takes less than 20 seconds.

Show HN: Cash: Tool for running shell commands on a large number of HPC nodes

lowsenberg — Tue, 13 Sep 2022 08:54:08 +0000

Article URL: https://github.com/janoliver/cash

Comments URL: https://news.ycombinator.com/item?id=32822001

Points: 5

# Comments: 2

New comment by lowsenberg in "Ask HN: Would you rather work 4 days per week for 80% salary?"

lowsenberg — Fri, 14 May 2021 17:19:26 +0000

I agree, I work part time (30hrs) at a company in Germany where "full time" is 35 hours/week, which more and more companies adopt here.

It's great. I'm done early in the afternoon and there is plenty of time left to spend with my family.

New comment by lowsenberg in "Processing 40TB of code from 10M projects with a dedicated server and Go"

lowsenberg — Tue, 01 Oct 2019 06:38:21 +0000

> The front-end varnish box for instance is doing the square root of zero most of the time.

Nicely written. :)

New comment by lowsenberg in "Ask HN: Physicists of HN, what are you working on these days?"

lowsenberg — Wed, 27 Mar 2019 14:46:56 +0000

I have a PhD in physics and worked as a post-doc for a few years, until I left for industry a couple of weeks ago. The last project I was busy with is developing a massively parallelized image simulation code for scanning transmission electron microscopy. It is open sourced here: www.stemsalabim.de

My new job in industry is consulting about HPC systems in the context of computer aided engineering.

New comment by lowsenberg in "A Daughter’s Disability and a Father’s Awakening"

lowsenberg — Sat, 12 Jan 2019 14:29:34 +0000

Your quotation marks seem like you disagree. There are other reasons not to do screening, for example not be faced with the decision to abort or not to abort. Here in Germany, at least in my circles, people tend to avoid such screenings not because of religiousness.

New comment by lowsenberg in "VisiData Lightning Demo at PyCascades 2018 [video]"

lowsenberg — Sun, 04 Mar 2018 18:35:41 +0000

Very neat, especially useful for peaking into (SQLite) databases and getting a quick overview of the contents. Thank you for your efforts!

New comment by lowsenberg in "TileDB: Storing massive dense and sparse multi-dimensional array data"

lowsenberg — Fri, 27 Oct 2017 10:55:50 +0000

This looks very interesting. Currently we are storing our dense simulation (and experimental) data in NetCDF/HDF5. Given correct chunking, this seems to be pretty efficient both performance and compression wise. What would we gain using TileDB? How does performance compare with HDF5?

New comment by lowsenberg in "Sublime Text 3.0"

lowsenberg — Wed, 13 Sep 2017 13:30:57 +0000

I don't know how often I type 'subl3 -n' every day. One of my all-time favorite tools. I'm happy to pay for the upgrade.

Generate PDF presentation with Inkscape and inkslides

lowsenberg — Sun, 20 Nov 2016 09:12:49 +0000

Article URL: https://github.com/janoliver/inkslides

Comments URL: https://news.ycombinator.com/item?id=12998018

Points: 5

# Comments: 0