Guillaume Dalle's Avatar

Guillaume Dalle

@gdalle.bsky.social

Researcher in machine learning and optimization. Open source enthusiast. Parody songwriter (aka PianoHamster). OCD survivor.

5,248 Followers  |  1,880 Following  |  600 Posts  |  Joined: 06.07.2023
Posts Following

Posts by Guillaume Dalle (@gdalle.bsky.social)

Best times for the NYT crossword by day of the week

Best times for the NYT crossword by day of the week

Crossword pro tip: solve it as a team with your smarter girlfriend. First time under an hour for a Sunday grid!

28.12.2025 16:44 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
a man in captain america 's uniform says i understood that reference Alt: captain america says i understood that reference
26.12.2025 12:25 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
ty: An extremely fast Python type checker and language server ty is an extremely fast Python type checker and language server, written in Rust, and designed as an alternative to mypy, Pyright, and Pylance.

I find it extremely comforting to know that Python development will basically be revolutionized every two years by @crmarsh.com and his crew at Astral.
astral.sh/blog/ty

17.12.2025 08:43 β€” πŸ‘ 35    πŸ” 3    πŸ’¬ 0    πŸ“Œ 0
Sarko, libre dans sa tΓͺte
YouTube video by PianoHamster Sarko, libre dans sa tΓͺte

youtu.be/pUNMqkv9oo4?...

11.12.2025 08:30 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Zotero
YouTube video by PianoHamster Zotero

youtu.be/jE83YERuwFQ?...

11.12.2025 07:44 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Post image

IJCAI 2026 will charge $100 USD per submission. Funds will be used to compensate reviewers.

2026.ijcai.org/ijcai-ecai-2...

04.12.2025 15:25 β€” πŸ‘ 28    πŸ” 4    πŸ’¬ 6    πŸ“Œ 11
My top song on Spotify this year was by β€œLes Fatals Picards”, I listened to it 235 times

My top song on Spotify this year was by β€œLes Fatals Picards”, I listened to it 235 times

Note to self: create a second Spotify account to work on song covers. Maybe this will allow my end-of-year recap to not look like an obsessive French pop fever dream.

03.12.2025 15:21 β€” πŸ‘ 6    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Awesome to see the French government recognize our work on open-source software! @ouvrirlascience.bsky.social

02.12.2025 19:03 β€” πŸ‘ 15    πŸ” 3    πŸ’¬ 1    πŸ“Œ 0
A close up of the trophy for the "prix science ouverte du logiciel libre de recherche, Γ©dition 2025" (open science prize for free research software, 2025 edition)

A close up of the trophy for the "prix science ouverte du logiciel libre de recherche, Γ©dition 2025" (open science prize for free research software, 2025 edition)

My baby #JuliaLang package got an award!
DifferentiationInterface.jl just received a national prize for open-source research software and I couldn't be prouder. We started this from a vague dream with @adrianhill.de, and reached 1000 dependents and 100k downloads last month.
Research != just papers.

02.12.2025 09:55 β€” πŸ‘ 89    πŸ” 13    πŸ’¬ 10    πŸ“Œ 1

The tune is "Shape of you"

30.11.2025 21:05 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

The ownership mechanism is a big one, but there are lots of small things with good design decisions: the packaging and build system, the module and file structure, the enum types. I’m also very excited about traits (because they’re a huge pain point in Julia) but I need to read further

30.11.2025 15:11 β€” πŸ‘ 8    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Trying to learn Rust for the Advent of Code, and literally every page of the manual has me going β€œwow, that *is* a problem I encountered in other languages and what a clever way to solve it”.

30.11.2025 13:35 β€” πŸ‘ 38    πŸ” 1    πŸ’¬ 3    πŸ“Œ 0

I take this to mean that you’re a scary insect

29.11.2025 22:36 β€” πŸ‘ 2    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0
Preview
Hype, Sustainability, and the Price of the Bigger-is-Better Paradigm in AI | Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency

Yes!!!
This is what prompted me to write the following paper:
dl.acm.org/doi/full/10....

A lot is explained by narratives and social norms driving research.
We need to build our own narratives, as a research community, this is what we do

29.11.2025 07:58 β€” πŸ‘ 8    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0

I have the title

28.11.2025 19:20 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I'm not sure how to interpret "across venues" here, which is very much load-bearing

28.11.2025 10:18 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Don't mind me, just going through old, irrelevant posts

28.11.2025 06:54 β€” πŸ‘ 20    πŸ” 1    πŸ’¬ 1    πŸ“Œ 0

each poster session at #neurips2025 has about 1000 posters. stupid question but isn't that too much?

27.11.2025 05:11 β€” πŸ‘ 7    πŸ” 1    πŸ’¬ 0    πŸ“Œ 0

Where do I buy this?

26.11.2025 12:12 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I sympathize very much, it definitely sucks. Not sure it helps, but there are doubts on the other side too: I often wonder whether academia is the right path for me. The best we can do is to make the choices that seem right for us one at a time, and it sounds like you did just that, so congrats πŸ«‚

26.11.2025 11:43 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Monotone operator theory be like: every algorithm is a fixed-point algorithm. A strangely comforting idea.

25.11.2025 10:26 β€” πŸ‘ 4    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

I’ve never seen something so ugly

24.11.2025 13:54 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 2    πŸ“Œ 0

I've been working with Zygote.jl for the last five years and I only just now got the name.

22.11.2025 08:34 β€” πŸ‘ 0    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Hell, if I were looking for a PhD position I’d jump at the chance too! But then again, who would supervise πŸ€”?

21.11.2025 18:39 β€” πŸ‘ 3    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

Link to job description: gdalle.github.io/assets/pdf/j...
[3/3]

21.11.2025 16:52 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

The location? Γ‰cole des Ponts, in the dreamy Paris suburbs.
The ideal applicant? A Master's student in mathematics or computer science, equally interested in theory, algorithms & open-source software.
Check out the offer, and shoot me an email if that sounds like you (or someone you know)!
[2/3]

21.11.2025 16:52 β€” πŸ‘ 4    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

⚠️ WANTED: PhD student ⚠️
The topic? GPU-friendly algorithms for discrete optimization and mathematical programming, to be used inside decision-focused learning pipelines.
The mentors? My colleague Axel Parmentier and myself.
[1/3]

21.11.2025 16:52 β€” πŸ‘ 14    πŸ” 4    πŸ’¬ 1    πŸ“Œ 1

Claude poets society

21.11.2025 00:01 β€” πŸ‘ 10    πŸ” 0    πŸ’¬ 0    πŸ“Œ 0

The Greek temple, cause it had so many columns

20.11.2025 05:33 β€” πŸ‘ 1    πŸ” 0    πŸ’¬ 1    πŸ“Œ 0

Teaching a class on decomposition methods for MILPs tomorrow and I'm wondering what the modern alternative is for "master problem" (as opposed to the pricing subproblem). Any suggestions? I've been using "main problem" but it feels too generic.

19.11.2025 12:49 β€” πŸ‘ 6    πŸ” 0    πŸ’¬ 5    πŸ“Œ 0