- x.com
- GitHubNew repoprereqs-of-deep-learning
Sasha Rush
ML researcher and blogger in New York; a professor at Harvard and then Cornell from 2015 to 2026, a researcher at Hugging Face from 2019 to 2024, and an ML researcher at Cursor, where he helped build the Composer models.
Sasha Rush did not write this page. ppll compiled it from their own words, each linked to the source. Is this you? · Report a mistake
Latest x.com
Feed
As its own page →Hiding
29 October
28 October
27 October
22 October
19 October
15 October
- x.com
- GitHubthe-token-game
- x.com
13 October
- GitHubjazz-piano
- x.com
8 October
1 October
15 September
7 July
13 April
1 January
20 June 2025
27 December 2024
18 November 2024
18 October 2024
7 October 2024
2 September 2024
Nothing matches.
Recommends rcmnd
rcmnd What they recommend 1 thing — each backed by an exact quote.
Compiled by rcmnd.app, not by them: the quotes are theirs, and each links to the post it was copied from.
Recent
Recommends video on distilling Mamba from Llama (“The Mamba in the Llama”)
RNN are neat. Here's a video describing how to make them work really well with little money:
x.com 28 Aug 2024
What others say about video on distilling Mamba from Llama (“The Mamba in the Llama”) →
Where they publish
Site rush-nlp.com Personal site: papers, talks and blog posts. Has a feed.
Link verified 20 Sept 2026.
x.com @srush_nlp Posts.
Recent
- Deep Learning Speedrun - (https://t.co/ePYCjVBrD6) What is the minimal amount of Math you need to train an MNIST model? 29 Oct 2026
- I've been really fascinated by the DL Alchemy course (https://t.co/4aSUZbOEs3). The decision to put the burden of theory construction on the student (and their claude) feels transformative and surprisingly challenging. Maybe knowledge in ML was never that critical. 28 Oct 2026
- Lean Verified Tokenizers (https://t.co/sX4zCF4C9y) Training-inference mismatch kills RL runs, but yet we build messages in Jinja? Prototypes a Lean-verified Rust agent pipeline. 27 Oct 2026
Show 5 more
- One feature of Vibed ML papers is that they do some cool hack in Section 3-4 and then Section 5 is called “Theoretical Analysis” as if we understood how any of this works. 22 Oct 2026
- Late to the party, but this blog post is excellent. Arguably the way people should first learn neural networks. 19 Oct 2026
- The Token Game (https://t.co/lSIzkGQWJr) It's BPE the mobile game? No reason for this to exist. I'm absolutely terrible at it. 15 Oct 2026
- Interpretability blogs are the recipes blogs of ML. Not sure, why I have to read through the story of your grandmother's first house to find out you used gradients. 15 Oct 2026
- How to Play Jazz Piano - https://t.co/PWMCH4edx1 A guide to jazz chords and harmonization with visualizations. (humanslop). 13 Oct 2026
Link verified 20 Sept 2026. Recent items update automatically from the channel.
YouTube @srush_nlp Talks and lecture recordings.
Link verified 20 Sept 2026.
GitHub @srush Code: the tensor and GPU puzzle series, and research repositories.
Recent
- New repoprereqs-of-deep-learning 29 Oct 2026
- the-token-game 15 Oct 2026
- jazz-piano 13 Oct 2026
Show 16 more
- jax-lean — Jax-to-Lean converter 8 Oct 2026
- sampling-to-reinforce 1 Oct 2026
- lean-transformer — Blog on formalizing key transformer invariants. 15 Sept 2026
- awesome-o1 — A bibliography and survey of the papers surrounding o1 7 Jul 2026
- g9jax 13 Apr 2026
- flashcard 1 Jan 2026
- annotated-s4 — Implementation of https://srush.github.io/annotated-s4 20 Jun 2025
- wgpu-shaders 27 Dec 2024
- osgai 18 Nov 2024
- Autodiff-Puzzles 18 Oct 2024
- colm-search 7 Oct 2024
- prof8 — Experimental paper writing linter. 2 Sept 2024
- GPU-Puzzles — Solve puzzles. Learn CUDA. 1 Sept 2024
- Tensor-Puzzles — Solve puzzles. Improve your pytorch. 15 Jul 2024
- MiniChain — A tiny library for coding with large language models. 10 Jul 2024
- DiffRast 1 Jul 2024
Link verified 20 Sept 2026. Recent items update automatically from the channel.
About the English under a post
Some people here publish in a language other than English. Where they do, this site shows a machine translation beneath the post, in this typeface — the site's own, not theirs.
The post itself is never changed, moved or hidden: what is set in the serif above is exactly what the person published, and it is what to quote them on. A translation can be wrong in ways that matter, especially about tone.
Only the post's own words are translated. A quoted post, a linked article and a belief on korrents.com are left in their original language.