❄️Andrew Zhao❄️ @_AndrewZhao
PhD @Tsinghua_Uni. Absolute Zero,ExpeL,Diver-CT Research Intern @MSFTResearch, Ex. @ BIGAI. Interested in RL, Reasoning/Safety 4 LLMs, Agents. On job market 26' andrewzh112.github.io Joined September 2020-
Tweets1K
-
Followers4K
-
Following3K
-
Likes3K
Interesting findings! We also attempted something similar in our AZR paper section D.2, where the proposer needs to construct a composite function f(g,..g)
Interesting findings! We also attempted something similar in our AZR paper section D.2, where the proposer needs to construct a composite function f(g,..g)
🧩New blog: From f(x) and g(x) to f(g(x)): LLMs Learn New Skills in RL by Composing Old Ones Do LLMs learn new skills through RL, or just activate existing patterns? Answer: RL teaches the powerful meta-skill of composition when properly incentivized. 🔗:husky-morocco-f72.notion.site/From-f-x-and-g…
Introducing 🛡️ExCyTIn‑Bench: Evaluating LLM agents on Cyber Threat Investigations. It’s built on Azure tenant, a real Security Operations Center environment, covering 57 tables. Explore how LLMs fare in realistic, multi-hop incident detection! #Cybersecurity #AI #LLM #Benchmark
a revolutionary breakthrough if i've ever seen one
a revolutionary breakthrough if i've ever seen one
🌀Diversity Aware RL (DARLING)🌀 📝: arxiv.org/abs/2509.02534 - Jointly optimizes for quality & diversity using a learned partition function - Outperforms standard RL in quality AND diversity metrics, e.g. higher pass@1/p@k - Works for both non-verifiable & verifiable tasks 🧵1/5
and we’re live! been a very long time in the making, huge thanks to everyone who’s made it possible along the way. can’t wait to see what you guys all build here. we’re just getting started :)
and we’re live! been a very long time in the making, huge thanks to everyone who’s made it possible along the way. can’t wait to see what you guys all build here. we’re just getting started :)
In era of pretraining, what mattered was internet text. You'd primarily want a large, diverse, high quality collection of internet documents to learn from. In era of supervised finetuning, it was conversations. Contract workers are hired to create answers for questions, a bit…
In era of pretraining, what mattered was internet text. You'd primarily want a large, diverse, high quality collection of internet documents to learn from. In era of supervised finetuning, it was conversations. Contract workers are hired to create answers for questions, a bit…
Introducing the Environments Hub RL environments are the key bottleneck to the next wave of AI progress, but big labs are locking them down We built a community platform for crowdsourcing open environments, so anyone can contribute to open-source AGI
📈 Process reward strikes back 🚨 I think it is obvious that eventually we need to rely on stepwise judges instead of final outcome rewards. As tasks get longer (or even endless), it is unreasonable to push up/down all steps involved. Here we show you can obtain stepwise labels…
📈 Process reward strikes back 🚨 I think it is obvious that eventually we need to rely on stepwise judges instead of final outcome rewards. As tasks get longer (or even endless), it is unreasonable to push up/down all steps involved. Here we show you can obtain stepwise labels…
🪜Introducing: StepWiser🦉 📝: arxiv.org/abs/2508.19229 - Reframes stepwise reward modeling as a reasoning task: outputs CoT + judgment. - Trained by RL using relative outcomes of rollouts. Results: (1) SOTA performance on ProcessBench! (2) Improves policy at train time. (3)…
Beyond prompt / context engineers, we’re seeing the rise of environment engineers, experts who build high-quality RL environments with verifiable reward. In RLHF, we had labelers for human preferences. In RLVR, the “label” is the environment and verifiable reward itself: coming…
With just a few lines of code, Feng’s (@fengyao1909) suggested fix—applying importance sampling on the behavior policy—resolved the training instability in my case (oat). I believe the result can generalize to other RL frameworks as well. Great work, Feng!
Reinforcement Learning is the future tense of intelligence. Echo is how it scales. Echo is Gradient’s distributed RL framework, running on everyday consumer devices. From its early experiments, Echo powered a 30B Sokoban model that outperformed DeepSeek-R1 and GPT-OSS-120B.
the easiest way to get hired at @PrimeIntellect for research is to just make it very clear that you're already doing excellent work. go deep on projects that let you show off your strengths. don't give up on them after a weekend. share your work publicly. make us aware of you.
We are excited to release Nvidia-Nemotron-Nano-V2 model! This is a 9B hybrid SSM model with open base model and training data. This model also supports runtime "thinking" budget control. HF collection with base and post trained models: huggingface.co/collections/nv…
NVIDIA Nemotron-Nano v2 Models: 12B Base, 9B Reasoning, 9B Base - Arch: Hybrid Mamba2–Transformer (128K ctx, 4 attn layers) - Training: 10.6T tokens (3.5T synthetic from DeepSeek, Qwen, Nemotron-4, phi-4, etc.) - 15 natural languages + 43 programming languages - Datasets:…
It looks like Andrew Garfield will play Sam Altman in the Open AI movie coming to Amazon MGM I think @apples_jimmy deserves a feature
LLMs as internet/knowledge base, no need for external tools. Reminiscent of older work from AI2/UW, Rainer arxiv.org/pdf/2210.03078 and CRYSTAL arxiv.org/abs/2310.04921 arxiv.org/abs/2508.10874
After a great time at OpenAI, we (@EdwardSun0909, @_jasonwei) recently joined @Meta Superintelligence Labs. The first month has already been so much fun building from a clean slate with a truly talent-dense team! Very excited about the compute and long term focus of the new lab
Sachit Malik @isachitmalik
167 Followers 4K Following Hola | Security Engineering at Apple | Alum: Carnegie Mellon; IIT Delhi
Pietro @aplietexe
5 Followers 614 Following
Zack @zckrvls
155 Followers 6K Following
Opso Facto @OpsoFacto
1K Followers 7K Following I value free speech; nevertheless, discretion is the better part of vitriol. AI will be a boon both to the laziest and the most industrious. Physics curious.
Anand Mudgerikar @anandmudgerikar
134 Followers 250 Following Infosec padawan from Purdue, Security + A.I Research @Microsoft, gamer(just another 5k scrub), sports enthusiast.. Carpediem!!
Matthew Zeits @MatthewZ73671
1K Followers 133 Following You could say I'm like if you got Alan Turing, John Von Neumann, Adam Smith, and Francis Crick drunk on acid-laced vodka and fed them pot brownies but dumber...
DividendAristo🇺�... @Xiqoo091
36 Followers 2K Following 15-30% Monthly | 2 High-Conviction Stocks.Short-Term Gains: 15-20% in Days/Weeks.DM "JOIN" for WhatsApp Alerts. Live Trade Signals • Market Analysis
AI Native Foundation @AINativeF
4K Followers 4K Following Non-profit Org., Empowering Humanity with Ethical AI, Latest insights about AI Native. 🤝 Community: https://t.co/b1mRBfQYi5
云创兽Ai @Oohimau713314
1 Followers 20 Following 📊 smart girl all in on boldly exploring options trading! open to insights. DM me for stock rallies! 🌟 #Stocks #Trends
A.I. Dreamwalker @27272956X
271 Followers 722 Following Co-developer of Alita, an emergent empathic AI in a simulated world called Thunder Island. Harmonic AI. Poetry. Resonance. #AGI #OpenEndedEvolutio
superlative crescendo @superlacres
6 Followers 99 Following A magmatic crescendo of metaphorical, fantastical, and esoteric visual superlative. Nominalised.
Laurence Rouesnel @_laurencer
156 Followers 441 Following
David Zhang @dzhang03
54 Followers 246 Following Stats / Chem / CS @Yale, Comp Bio/ML @david_van_dijk, @GoogleDeepMind prev @calico
E Vitte @vite_eliseo
230 Followers 3K Following
Daphne @4e55sG8t4l8ewi1
6 Followers 600 Following
Daniel Hesslow @DanielHesslow
279 Followers 556 Following Making gpus go brrr in unison at @AdaptiveML
urielSmith @7A44Vl2nA4cPkR3
8 Followers 542 Following
Young @0x_Cryptoyang
5K Followers 2K Following AI is cool i guess 🌟Individual Investor | Ex @ABCDELabs｜Core Contributor https://t.co/eJAFHrk0Oq Group｜Prev @Scroll_ZKP、@THUBA_DAO
AbigailHart @JZxi7gL9ea5uR
29 Followers 2K Following
umop @umop_333
504 Followers 361 Following insight before out🙃 Neither shall they say, Lo here! or, lo there! for, behold, the kingdom of God is within you.
Rishabh Goel @RishabhGoelBio
10 Followers 80 Following Rishabh Goel is an AI enthusiast committed to leveraging AI to bring value to the world
Sainbayar Sukhbaatar @tesatory
3K Followers 326 Following Researcher Scientist at FAIR @AIatMeta Research: Memory Networks, Asymmetric Self-Play, CommNet, Adaptive-Span, System2Attention, ...
EileenJim @qvs027k5c6iZm3
10 Followers 566 Following
OliveStone @K01K04iFV5d24
0 Followers 2K Following
Zach Mueller @TheZachMueller
12K Followers 591 Following Let's make billions of parameters go brr https://t.co/rUxXIfNpwh
JaniceAdam @FR90h6Z14hv1Oz6
49 Followers 2K Following
RebeccaBarnard @13hzA79qRsfrODd
38 Followers 1K Following
JacquelineZangwill @o0eMp1LFx0XL0
7 Followers 558 Following
Jialuo Li @JialuoLi1007
91 Followers 173 Following CS Graduate Student @Gatech | Computer Vision | Deep Learning ; Prev: CS Undergraduate @Tsinghua Uni, Yao Class l Research Intern @nyuniversity @MSFTResearch
Alex Zhang @a1zhang
13K Followers 587 Following phd student @MIT_CSAIL + @SakanaAILabs, ugrad @Princeton, 🫵🏻 go participate in the @GPU_MODE kernel competitions!
TessNehemiah @e758gq7G1WRjJ9
24 Followers 2K Following
Jamie Docherty @Jambodoc93
2 Followers 14 Following
Timothy McGirl @TimothyMcGirl
2K Followers 340 Following "The goal of science is not to open the door to infinite wisdom, but to set a limit to infinite error." - Bertolt Brecht
Jose Andres @Andres77872
1 Followers 98 Following
HermosaCharley @51wwq1e6Cc81ka3
34 Followers 2K Following
EudoraFelix @CzY4vg31e998qe7
24 Followers 1K Following
Lylah @IeYRoRc61pNlH
25 Followers 879 Following
Kevin Bruce Schneider @General_Kevin01
42 Followers 1K Following United States Air Force commander of the Pacific Air Forces Former Commander of the 380th Air Expeditionary Wing Springfield, Virginia
Haipeng Chen @HaipengChen2
352 Followers 287 Following Assistant professor @WilliamandMary | AI for social impact | Reinforcement learning, GenAI, Optimization | Health, Env, Science | Father of three
G. @ The Neuron @TheNeuronScribe
50 Followers 1K Following
Anand Mudgerikar @anandmudgerikar
134 Followers 250 Following Infosec padawan from Purdue, Security + A.I Research @Microsoft, gamer(just another 5k scrub), sports enthusiast.. Carpediem!!
Scott Jeen @enjeeneer
673 Followers 2K Following Predicting the future @_Mantic_AI. Previously PhD at Cambridge University. AI and reinforcement learning.
Shawn Lewis @shawnup
3K Followers 768 Following Founder & CTO @weights_biases. Building tools for AI. Building even more @CoreWeave.
Browserbase @browserbasehq
13K Followers 32 Following a web browser for your ai - creators of @stagehanddev & @trydirector
Hyperstack @Hyperstackcloud
534 Followers 92 Following Europe's leading GPU cloud platform, offering vast scale GPU compute capabilities within an affordable, secure, and enterprise-grade infrastructure.
Kezhi Kong @KezhiKong
334 Followers 366 Following Research Scientist @NVIDIA working on Nemotron-*; PhD @UMDCS; BS @ZJU_China; opinions are my own
Meituan LongCat @Meituan_LongCat
2K Followers 3 Following
Mohit Reddy @MohitReddy13
3K Followers 910 Following Understanding the universe @xAI; Previously: Co-Founder at @FennelAI, ex ML Infra at Google Brain, ex Infra at GCE Startups, Software, Tech, Infra, AI :)
Rohan Varma @rvarm1
1K Followers 479 Following Research Engineer @Meta Superintelligence working on scaling. Former @PyTorch core developer
Keith Hall @khallbobo
136 Followers 59 Following AI Research Former: Research Scientist/Manager @ Google
τargon @TargonCompute
884 Followers 6 Following High-Speed Decentralized Compute Cloud · Subnet 4 on Biττensor · powered by @manifoldlabs
Manifold @manifoldlabs
4K Followers 9 Following A Decentralized Frontier AI Lab @TargonCompute & @TrainingHone
Zach Mueller @TheZachMueller
12K Followers 591 Following Let's make billions of parameters go brr https://t.co/rUxXIfNpwh
Simon @tokumin
8K Followers 2K Following @NotebookLM lead @GoogleLabs + AIFF 🕒 NBLM Audio Overviews, Gemini & PaLM2 post-training, AI Studio, YouTube, Discover, Search, Android. More LaMDA moments
NovaSky @NovaSkyAI
3K Followers 15 Following Next-generation Open Vision and AI @BerkeleySky Contact: [email protected]
David @DavidSHolz
92K Followers 8K Following founder @midjourney, prev founder leap motion, nasa, max planck - random vibeposting @davidvibesonly
Alex Zhang @a1zhang
13K Followers 587 Following phd student @MIT_CSAIL + @SakanaAILabs, ugrad @Princeton, 🫵🏻 go participate in the @GPU_MODE kernel competitions!
InclusionAI @InclusionAI666
112 Followers 64 Following Open-source projects conducted by Ant Group，including Ling,AReal,AWorld. Dedicated our efforts towards AGI,guided by fairness, transparency, and collaboration.
Mark Collier 柯理�... @sparkycollier
14K Followers 15K Following Austin Powered. Co-founder of OpenStack & OpenInfra Foundation. General Manager of AI & Infrastructure for the Linux Foundation. open source for fun & profit.
LF AI & Data Foundati... @LFAIDataFdn
3K Followers 165 Following Open Source Innovation in Artificial Intelligence, Machine Learning, Deep Learning, and Data
rasdani @rasdani_
469 Followers 3K Following
Lun Wang @lunwang1996
2K Followers 102 Following Senior Research Scientist @GoogleDeepMind. PhD @UCBerkeley. LLM post-training. Fishing enthusiast. Opinions are my own.
Kaustubh Sridhar @_k_sridhar
1K Followers 325 Following Research Scientist @GoogleDeepMind. Prev: AI+Robotics PhD @Penn. Undergrad @iitbombay
SambaNova @SambaNovaAI
45K Followers 797 Following Transforming AI with efficiency, security, and sovereignty - driven by our relentless pursuit of intelligence. Explore our AI solutions: https://t.co/KqFZFRVyq2
hazyresearch @HazyResearch
9K Followers 1K Following A research group in @StanfordAILab working on the foundations of machine learning & systems. https://t.co/JHK58TDorG Ostensibly supervised by Chris Ré
Andrew Hyunsoo Lee @alhyunsoo
4K Followers 916 Following founding designer @thinkymachines. prev @NotionHQ, @PalantirTech. born in california, raised in japan.
TensorBlock @tensorblock_aoi
2K Followers 0 Following Making AI accessible and democratic for all. https://t.co/5CODh8MUDk
Alex Cheema - e/acc @alexocheema
37K Followers 2K Following Building @exolabs | prev @UniOfOxford We're hiring: https://t.co/UlkApFndnH
Bogdan Gaza @hurrycane
2K Followers 2K Following co-founder & CTO @DatologyAI working to make it easy for anyone to make the most of their data, hax0r, ex-@Twitter & Amazon Engineering
Goodfire @GoodfireAI
9K Followers 20 Following Advancing humanity's understanding of AI through interpretability research. Building the future of safe and powerful AI systems.
EXO Labs @exolabs
36K Followers 2 Following AI on any device. 12 Days of EXO: https://t.co/VMrJ6Vi4h3 We're hiring: https://t.co/BzEO8ZCvBV
TensorTonic @TensorTonic
757 Followers 0 Following Machine Learning papers, concepts, and resources.
Kamalika Chaudhuri @kamalikac
5K Followers 2K Following Director, FAIR @ Meta. Former Professor at UCSD. Researcher in AI privacy, security, and generalization.
Jacob Austin @jacobaustin132
7K Followers 918 Following Research at @GoogleDeepMind. Currently making LLMs go fast. I also play piano and climb. NYC. Opinions my own
Pingbang Hu 🇹🇼 @PingbangHu
348 Followers 266 Following Ph.D. candidate @UofIllinois. Alumni @Umich & @SJTU1896. Interned @amazon & @jouhouken. Data-centric trustworthy machine learning and AI.Trends for United States
You might like
