The PDF version of the (very early version of) Machine Learning Engineering Open Book is ready for your enjoyment. Download it from here: github.com/stas00/ml-engi… If you find any problems please kindly let me know. Thank you Julien Chaumond for giving me permission to host it on

Mamba apparently was rejected !? (openreview.net/forum?id=AL1fq…) Honestly I don't even understand. If this gets rejected, what chance do us 🤡 s have.

In other news; I’m joining Aleph Alpha where among other things, I will continue my work on building trustworthy and efficient foundation models and contribute towards building sovereign AI. Very excited for this new challenge 😄

I’m incredibly grateful to have found an advisor like John who’s a great researcher but an even better person. There were times in my PhD I would’ve quit had it not been for his unconditional support. Academia is lucky to have him!

if you want to get more familiar with various ways of splitting a matmul take a look at the little visualization i wrote over the weekend

maharshi At one point a while back autoregressive language model papers were like that too. Formulating the joint likelihood, factorizing it, deriving the maximum likelihood estimate, discussing connections to Bayesian statistics and Convex Optimization,... Good example here:

👋I am on the academic job market, looking for tenure-track positions. I work on machine learning, decision making, and social aspects of AI. Let's get in touch if your institution is hiring! 💻stsirtsis.github.io 😀Shares are very much appreciated!

severely underrated playlist, will binge this

Highly recommend watching Tom Goldstein's talk: Using recurrence to achieve weak to strong generalization youtube.com/live/M7Kq0ooFF…

Excited for ICLR 2025 in Singapore? Join our BoF Social (24 Apr, 12:30 p.m., Opal 103-104) on tokenizer-free, end-to-end architectures. Ready for insightful discussions and networking? Sign up here forms.office.com/e/UaffUVtyHx #ICLR2025 #AIResearch #EnterpriseAI #Tokenizers

If you are at #ICLR25 and care about tokenizers, drop by our (Aleph Alpha)’s Birds of a Feather session – happening now at Opal 103.

For #ICLR2025 we are unveiling a new, high-quality pretraining dataset for German LLMs. Shared to strengthen the open research community. Shaped by our belief in excellence and transparency. huggingface.co/datasets/Aleph…