In France STEM PhD are expected to last 3 year and the funding is sized like that. It is also considered as a job. It is only done if salary is funded in most cases.

Often it spill over a bit and I guess France travail (French job agency managing insurance for people losing their job) should often be cited /thanked in Phd student thesis for funding the final steps of their manuscript.

There are limited internship culture during the phd itself Afaik.

However phd is never started at Bachelor level, only after Msc that last two years and requires an internship or research projet.

I heard a person saying a bit like you that it is not enough to grow a Real expert though compared to US phd. But Oftentimes postdoc always follow for Longer and longer

I use docker compose a lot for small projects on a single box like Ovh vps. With a reverse proxy on the host. Nice ans easy to setup and update.

If you want to test something that is between compose and k8s, check ring: https://github.com/kemeter/ring

Did you market the project as trusted? Récent update (à few month) requises the trust to reenable the analyses feature It took me a while to understand lol At Somme point I though that the parker were broken in my codebase xD

I did not. Ty. I will look into doing this.

You might be interested by tournesol.app

Deezer has been quite stable imo and I find it not clutered

No linux client and support in my country. They have better streaming quality than Spotify.

Can you tell us a bit more about this creative residence and how did you get accepted there ?

Yes, it's called Almost Perfect. I previously followed the illustrator Luis who co-runs it, and applied back in 2022.

https://almostperfect.jp

However, the current "version" of Almost Perfect is ending in about a year, and I think they are fully booked until then:

https://www.instagram.com/almostperfecttokyo/p/C2PCMtpS_wq/

It's been a fun experience, though - I shared a house with an illustrator from LA. I spent time exploring Tokyo and working on Booklet. At the end, every resident gives a gallery - normally it's visual, such as drawings, but mine became a presentation.

Young machine learning practitioners : That is why you should do not trust accuracy when doing a ML model on imbalanced datasets.

And in general, better use precision, recall, f1-score or confusion matrices

I quite like the idea of being skeptical of something but still hedging on its potential success. Thanks

The best way imo is the telegram bot

hey very nice work !

do you think that a docker.js running containers in the browser is something that could become a reality soon if we follow you line of work ?