news | Maximilian Beck

</table> </div> </div>

Apr 13, 2026	I have started a new position as AI Research Scientist at Meta FAIR Paris. I will continue my work on efficient Language Model architectures.
Mar 19, 2026	I have successfully defended my PhD thesis with title xLSTM: Recurrent Neural Network Architectures for Scalable and Efficient Large Language Models Special thanks to Sepp Hochreiter for his excellent guidance throughout my PhD. (Pre-Recording) (Slides) (PhD-Thesis)
Mar 10, 2026	We have published our paper Towards a Neural Debugger for Python on arxiv. I had a lot of fun building this during my internship at Meta FAIR Paris last summer. Huge thanks to mentors and collaborators Jonas Gehring, Jannik Kossen and Gabriel Synnaeve! 🧠🪲We introduce Neural Debuggers: 🧑‍🏭 LLMs that emulate traditional debuggers by predicting forward code execution (future states & outputs) and inverse execution (inferring prior states or inputs) conditioned on debugger actions such as step over, step into, or breakpoints. pic.twitter.com/rucET8GFBA— Maximilian Beck (@maxmbeck) March 11, 2026</blockquote> </td> </tr>
Jan 22, 2026	Our two papers xLSTM Scaling Laws: Competitive Performance with Linear Time-Complexity and Short window attention enables long-term memorization are accepted at ICLR 2026 in Rio de Janeiro!
Oct 24, 2025	I have successfully finished my internship at Meta FAIR with Jonas Gehring. I am happy to have contributed and co-authored two papers, with one more in preparation - Stay tuned!
Oct 03, 2025	We published our paper xLSTM Scaling Laws: Competitive Performance with Linear Time-Complexity on arxiv. 🚀 Excited to share our new paper on scaling laws for xLSTMs vs. Transformers. Key result: xLSTM models Pareto-dominate Transformers in cross-entropy loss. - At fixed FLOP budgets → xLSTMs perform better - At fixed validation loss → xLSTMs need fewer FLOPs 🧵 Details in thread pic.twitter.com/Rd3s08BtQY— Maximilian Beck (@maxmbeck) October 3, 2025</blockquote> </td> </tr>
Sep 06, 2025	I have handed in my PhD pre-defense report for my PhD Thesis: xLSTM: Recurrent Neural Network Architectures for Scalable and Efficient Large Language Models
May 12, 2025	I started a summer internship at Meta FAIR in Paris working on code Large Language Models.
May 10, 2025	Our two papers xLSTM 7B and A Large Recurrent Action Model are accepted at ICML 2025!
Apr 28, 2025	Two main and two workshop papers at ICLR 2025.
Jan 22, 2025	Two papers accepted at ICLR 2025! Our FlashRNN and Vision-LSTM paper will be presented in Singapore.
Sep 25, 2024	🚨 The xLSTM paper has been accepted as a spotlight at NeurIPS 2024! 🎉 I am looking forward to present and discuss the xLSTM in Vancouver.
Jul 26, 2024	We presented the xLSTM at 3 Workshops at ICML 2024 including an Oral at ES-FOMO@ICML2024.
Jul 01, 2024	I presented the xLSTM on 1littlecoder's YouTube Channel.
Jun 27, 2024	ELISE wrap up conference 2024 in Helsinki. I presented the xLSTM as ELLIS PhD Spotlight presentation.
May 08, 2024	It's out! We published the xLSTM on Arxiv.
Apr 08, 2024	My fourth PhD Talk presenting first results of the xLSTM.
Oct 30, 2023	My third PhD Talk presenting the details of the xLSTM for the first time.
Aug 28, 2023	ELLIS Doctoral Symposium 2023 in Helsinki. I presented our ICLR23 paper Addressing Parameter Choice Issues in Unsupervised Domain Adaptation by Aggregation.
Nov 28, 2022	My first big Machine Learning conference: NeurIPS 2022 in New Orleans.
Nov 21, 2022	My second PhD Talk in our Seminar about “Loss Landscapes under Distribution Shift”. Slides.
Oct 16, 2022	My first invited talk together with Martin Gauch about SubGD: Few-Shot Learning by Dimensionality Reduction in Gradient Space at Ruhr-University Bochum.
Sep 19, 2022	ELLIS Doctoral Symposium 2022 in Alicante. I presented SubGD.
Aug 22, 2022	CoLLAs 2022 in Montreal. Together with Martin Gauch, I presented our Few-shot Learning method SubGD.
May 16, 2022	My first PhD Talk in our internal PhD Seminar about SubGD: Few-Shot Learning by Dimensionality Reduction in Gradient Space. Slides.