PhD student at @Uni_Stuttgart @MPI_IS @ELLISforEurope | ML Researcher at @Bitdefender | https://xitter.cf/t.co/ljYnqFqJgZ

Joined July 2016
I'll be at #ICML2026 in Seoul next week presenting three works with some great collaborators. Quick rundown in this thread. Come say hi if you're around, always happy to chat!
2
3
24
1,518
The takes on Terry Tao being unemployed seem to me to be an expression of mediocrity and resentment, from people who have never tasted dedication to something higher than themselves. The guy is not going to lose his job. He is forever in the Hall of Fame. If he was purely selfish, he'd be grinning that he got on the train right before it crumbled. Maybe, just maybe, he's sad to see something he cared about potentially crumble, instead of being self- interestedly afraid "for his job".
47
128
9
1,672
58,962
My LLM benchmark is how ridiculous it can make some themes for my personal website. Old ones made with ChatGPT 5.5 and Opus 4.6 were ok but kinda slop. Opus 5.5 is peak. Even added ascii art and background sounds lol (before->after screenshots)
4
242
Andrei Manolache retweeted
And thatโ€™s a wrap on Romanian AI Days 2026! ๐Ÿ‡ท๐Ÿ‡ด๐ŸŽ‰ A heartfelt thank-you to everyone who made Romanian AI Days 2026 possible: our speakers, sponsors, organizers and attendees! You made this edition an unforgettable milestone! ๐Ÿค— Here are the key takeaways from Day 2 ๐Ÿงต:
8
2
11
286
Andrei Manolache retweeted
book tour continues la politehnica ๐Ÿ‡ท๐Ÿ‡ด
2
45
3,542
Next stop: #ECCV2028 Bucharest, Romania ๐Ÿ‡ท๐Ÿ‡ด
3
22
3
196
12,658
Every time you refresh, OpenReview adds +1 minute to the wait before your #NeurIPS reviews drop.
1
6
2,675
Why are workshops usually after the main conf? ML conferences got so big that it can be a struggle to find people interested about the same topics. Workshops feel like the only place where this is still possible, would be nicer to have them before the main conf @icmlconf et al
4
180
Andrei Manolache retweeted
Happy that GraphBench has been accepted as an oral at the GFM workshop @ICML. ๐Ÿš€ We also updated GraphBench๐Ÿ†•: - Extensive documentation๐Ÿ“”, - more baselines๐Ÿ“Š, - bugfixes๐Ÿชฒ, - extended 80+ page preprint ๐Ÿ“ฐ(arxiv.org/abs/2512.04475). Check it out: graphbench.io/.
8
22
2,845
How the Optimizer Shapes Learned Solutions in Equivariant Neural Networks, at the ICML WSS Workshop. Swapping optimizers changes the solutions of equivariant nets, down to the loss geometry and spectra. With Teodor Stupariu, who did this as a highschool intern at @Bitdefender ๐Ÿคฏ
1
1
1
186
arXiv links: TEDBench: arxiv.org/pdf/2605.18552 C-FREE: arxiv.org/pdf/2509.22468 Optimizer + Equivariance: arxiv.org/pdf/2605.27662 See you in Seoul! ๐Ÿ‡ฐ๐Ÿ‡ท
1
71
I was starting to wonder if I was getting rusty, it's been a while since I received one of these. Thanks for the recognition, @icmlconf! ๐Ÿซก
1
10
613
Andrei Manolache retweeted
โฐ The CoLLAs abstract deadline is only 10 days away! We invite researchers to explore all facets of ML adaptation, from incorporating new capabilities during continuous training to efficiently removing outdated or harmful data. - ๐—”๐—ฏ๐˜€๐˜๐—ฟ๐—ฎ๐—ฐ๐˜ ๐——๐—ฒ๐—ฎ๐—ฑ๐—น๐—ถ๐—ป๐—ฒ: April 10, 2026 - ๐—ฆ๐˜‚๐—ฏ๐—บ๐—ถ๐˜€๐˜€๐—ถ๐—ผ๐—ป ๐——๐—ฒ๐—ฎ๐—ฑ๐—น๐—ถ๐—ป๐—ฒ: April 15, 2026 - ๐—–๐—ผ๐—ป๐—ณ๐—ฒ๐—ฟ๐—ฒ๐—ป๐—ฐ๐—ฒ ๐——๐—ฎ๐˜๐—ฒ๐˜€: Sep 14โ€“17, 2026 ๐Ÿ“š Accepted papers will be published in the Proceedings of Machine Learning Research (PMLR). ๐Ÿ”— ๐—™๐—ผ๐—ฟ ๐—ณ๐˜‚๐—น๐—น ๐—ฑ๐—ฒ๐˜๐—ฎ๐—ถ๐—น๐˜€ ๐—ผ๐—ป ๐˜๐—ต๐—ฒ ๐—–๐—ฎ๐—น๐—น ๐—ณ๐—ผ๐—ฟ ๐—ฃ๐—ฎ๐—ฝ๐—ฒ๐—ฟ๐˜€: lifelong-ml.cc/Conferences/2โ€ฆ
6
2
16
8,058
Andrei Manolache retweeted
Weโ€™re bringing the CoLLAs 2026 community to Bucharest, Romania ๐Ÿ‡ท๐Ÿ‡ด next September! ๐Ÿง  Focus: Adaptation, Continual Learning, Unlearning ๐Ÿ“… Submit by mid-April Chat about CoLLAs topics or must-see spots in Bucharest at #NeurIPS2025 with Sarath, Razvan, and Eleni! ๐Ÿš€
The CoLLAs 2026 website is LIVE! We invite researchers to explore all facets of ML adaptation, from incorporating new capabilities during continuous training to efficiently removing outdated or harmful data. ๐Ÿ“ Location: Bucharest, Romania ๐Ÿ“… Date: Sep 14โ€“17, 2026 ๐Ÿ”— Read the full Call for Papers: lifelong-ml.cc/Conferences/2โ€ฆ Sarath Chandar, Razvan Pascanu, and Eleni Triantafillou are currently at NeurIPS and happy to chat about CoLLAs.
3
12
1,068
Andrei Manolache retweeted
I respect that @iclr_conf had to respond to the OR leak, but I disagree with resetting scores. Many students worked hard on rebuttals and improved their papers in good faith. I hope the organizers reconsider and revert the reset. If you agree, feel free to retweet.
6
45
3
149
20,316
Andrei Manolache retweeted
ChronoGraph (w/ A. Lutu, I. Pintilie & @ilarele) is a real-world dataset of microservice telemetry that links multivariate time series with a dependency graph and incident labels, enabling evaluation of forecasting and anomaly detection methods under propagation effects. ๐Ÿงต8/9
1
2
1
146
Andrei Manolache retweeted
Inspired by JEPA, C-FREE (w/ B. Ariguib & @mniepert) uses a predictive objective over egonets spanning 2D and 3D graphs, avoiding negatives, reconstructions, and augmentations. This simple multimodal design delivers strong results from low- to large-scale regimes. ๐Ÿ”ฌ ๐Ÿงต6/9
1
2
4
276