Shashank Gupta @shashank_bits
Researcher at Ai2 || Work on NLP, LLMs, Reasoning, Agents, AI4Code || Prev: Microsoft AI, Univ. of Illinois (UIUC), Max Planck, IIT-Bombay shashankgupta.info Seattle, WA Joined December 2010-
Tweets605
-
Followers611
-
Following2K
-
Likes7K
The first autonomous agent cyberattack is an unprecedented event that deserves unprecedented transparency. Today we’re sharing everything we can: a full technical timeline, an interactive replay, and how we used an open model to defend ourselves, so defenders everywhere can learn from it and prepare for what’s next. huggingface.co/blog/agent-int…
We've open-sourced AgentENV in collaboration with kvcache-ai. AgentENV is a distributed system for running agent environments at scale. Its components power agentic RL training for Kimi K3, with fast snapshot, resume, and fork support for large-scale parallel agent workflows. Explore on GitHub: github.com/kvcache-ai/Age…
A super long overdue (3+ years?) post on scaling laws. Compute is expensive. Scaling laws are a way to help us reason about the optimal compute allocation between data and model size before committing to a large run. The post covers what scaling laws predict, how compute-optimal allocation works, why Kaplan et al. and Chinchilla disagree, and how data limits + fitting details make extrapolation tricky. lilianweng.github.io/posts/2026-06-…
Trained some terminal agents with friends! Introducing Tmax, open RL terminal agent models. Under default settings and shorter length (65k) token budgets, tmax outperforms prior open work on terminal use. We are releasing all data+weights+rollouts publically!
@YueYangAI @allen_ai @thinkymachines Congrats, and good luck, Yue! ✨🚀
ECHO is now on arxiv. Please share your thoughts and comments!
🚀 DeepSeek-V4 Preview is officially live & open-sourced! Welcome to the era of cost-effective 1M context length. 🔹 DeepSeek-V4-Pro: 1.6T total / 49B active params. Performance rivaling the world's top closed-source models. 🔹 DeepSeek-V4-Flash: 284B total / 13B active params. Your fast, efficient, and economical choice. Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today! 📄 Tech Report: huggingface.co/deepseek-ai/De… 🤗 Open Weights: huggingface.co/collections/de… 1/n
Train LLMs to manage their own cache directly with rl!! No summarization in tokens, direct evictions of blocks trained end to end with RL. Very cool work from @michaelyli__
cool new paper from @michaelylin, @jubayer_hamid, @efoxdcs and @noahdgoodman! instead of sparse attention or heuristic evictions, they RL a model to evict from its own KV cache, learning to manage its own memory
@MKhalifaaaa @NVIDIAAI @YejinChoinka @rajammanabrolu @aviral_kumar2 Congrats, Muhammad!!! 🎉😀
1/ It’s been so fun working with @shengjia_zhao, @alexandr_wang and the team to build muse spark from scratch. It is early and has rough edges, but excited to continue our research velocity. I especially love that we’re doubling down on the fundamental science. We’re focused on building methods that scale well along all axes: pretraining, RL, and test-time compute. ai.meta.com/blog/introduci…
If you want to understand the latest landscape of RL training and frameworks this blog by @DirhousssiAmine and the @huggingface team compares 16 different RL frameworks from VeRL, SLIME , TRL to many more across the following aspects. > Orchestration & Concurrency Primitive: + how distributed components are coordinated (Ray actors, asyncio, pub/sub, HTTP). > Rollout Buffer Design: + how rollouts flow from inference to training. > Weight Synchronisation Protocol: + how updated weights reach inference servers, and whether the system must pause to accept them or continue generating. > Staleness Management: + how off-policy rollouts are handled: version rejection, depth bounding, or importance-sampling correction. > Partial Rollout Handling: + what happens to in-flight generations when a weight update arrives mid-sequence. > LoRA Training Support: + General LoRA support and whether adapter-only parameters can be trained and synced, enabling sub-millisecond weight transfers. > Distributed Training Backend & Parallelism: + what parallelism strategy is used for training, constraining max model size. Its very well written and i was able to learn a lot from it!
This is all overblown. Yes some people are leaving, no Ai2 is not done developing open models. x.com/xlr8harder/sta…
A tragedy is unfolding at AllenAI. Fs in the chat
New paper from our team is now on arxiv. One of the techniques we use for Nemotron-3 post-training. "PivotRL: High Accuracy Agentic Post-Training at Low Compute Cost"
We're releasing a technical report describing how Composer 2 was trained.
Today we're releasing MolmoWeb, an open source agent that can navigate + complete tasks in a browser on your behalf. Built on Molmo 2 in 4B & 8B sizes, it sets a new open-weight SOTA across four major web-agent benchmarks & even surpasses agents built on proprietary models. 🧵
@vjhofmann @allen_ai @CisLmu @LMU_Muenchen Congrats, Valentin! 🥳🎉
Archiki is amazing!! ✨ Hire her! 💼
🚨 I’m on the 2026 Research Scientist Job Market! I am a PhD student at UNC Chapel Hill (advised by @mohitban47) and recipient of the Apple Scholars in AI/ML PhD Fellowship. My research centers around: 🔸Reasoning & RL/Post-Training: Evaluating and interpreting the reasoning
Introducing Theorizer: Turning thousands of papers into scientific laws 📚➡️📜 Most automated discovery systems focus on experimentation. Theorizer tackles the other half of science: theory building—compressing scattered findings into structured, testable claims. 🧵
Luca Soldaini 🎀 @soldni
13K Followers 1K Following data mines are my passion ⛏️ mts @MicrosoftAI / ex Olmo co-lead @allen_ai / posts mine, leave my employer alone / 🌈
Swaroop Mishra @Swarooprm7
15K Followers 765 Following Research Scientist at @GoogleDeepmind, Opinions mine
Akari Asai @AkariAsai
23K Followers 949 Following Assistant Professor @SCSatCMU & research scientist @allen_ai
Sarah Wiegreffe @sarahwiegreffe
6K Followers 1K Following Research in NLP & trustworthy ML (mostly interpretability). Trying to make LMs more transparent, reliable, & safe. Assistant prof @umdcs @clipumd Views my own.
Faeze Brahman @faeze_brh
2K Followers 1K Following Sr. Research Scientist @allen_ai | Prev. Postdoc @allen_ai @uwnlp | Ph.D. from UCSC | Former Intern @MSFTResearch , @allen_ai | Researcher in #NLProc, #ML #AI
Yao Fu @Francis_YAO_
23K Followers 2K Following Prev. Grok 4.5/4.6 pretraining scaling @xAI; Gemini 3 perception and project Astra @GoogleDeepMind
Matthew Finlayson @mattf1n
1K Followers 940 Following @GoodfireAI fellow, PhD at @nlp_usc | Former predoc at @allen_ai on @ai2_aristo | Harvard 2021 CS & Linguistics
Nouha Dziri @nouhadziri
6K Followers 696 Following Researcher @cohere, ex @allen_ai, @GoogleDeepMind @MSFTResearch @MilaQuebec 🚨 PhD in LLMs 🤖 UofA. my blogs about LLMs reasoning: https://t.co/zJCt9jdYzJ
Weijia Shi @WeijiaShi2
11K Followers 2K Following @uwnlp @aclmentorship | Prev @allen_ai @MetaAI @CS_UCLA
Valentina Pyatkin @valentina__py
3K Followers 2K Following ETH AI Center, previously @allen_ai , @uwnlp
Yizhong Wang @yizhongwyz
7K Followers 2K Following Researcher at ByteDance Seed, Incoming Asst Prof @UTAustin. Agent training, agents for science discovery.
Harsh Trivedi @harsh3vedi
746 Followers 932 Following Research Scientist @allen_ai 🤖 building AI agents & environments: 🌍 AppWorld (https://t.co/dIawTLcI7a) Prev: #NLProc PhD @stonybrooku. On 🦋 same handle.
Eric Zelikman @ericzelikman
24K Followers 2K Following cofounder & ceo @humansand - building ai for humans // was lgtm-ing @xAI, phd-ing @stanford
Siddarth @siddarthv66
2K Followers 926 Following Post-training @MistralAI, PhD @Mila_Quebec | RLing LLMs
Amy Latina 🔥 @amybliss6903jj
12 Followers 249 Following this is your motion 🫦 watch my main: @amyblissxo
Simran Khanuja @simi_97k
4K Followers 1K Following PhD @LTIatCMU | Prev. @GoogleResearch @MSFTResearch
Abel Le @tasuke2k3
94 Followers 5K Following
Hamsa @Hamsa02309
114 Followers 527 Following low-latency | showing up | getting things done | omw to @openai
Hazel @MahajanHazel
194 Followers 2K Following SDE 2 @Cisco MS CS @ubuffalo , Ex @rfsuny , @Delhivery, @revolaAI 25👩🏻💻, Open to tech Collabs Let's chat - https://t.co/dZpkrqggHT
Gautam @Gautam2086
351 Followers 8K Following full-stack ai engineer | ai agents & ml infra | 1,260+ leetcode (top 9%) | python, typescript, aws | ex- @rfsuny, @ChargePointnet, @blox_xyz, @BioProtocol
Zijian Chen @zijian42chen
265 Followers 150 Following Information Retrieval & NLP @uwaterloo, Intern @allen_ai, "Steven".
murali @muraliai1
2 Followers 32 Following
Joachim Baumann @joabaum
900 Followers 2K Following Postdoc @StanfordNLP @StanfordAILab / Prev: @MilaNLProc @UZH_en @MPI_IS @CarnegieMellon. CompSocSci, LLMs, algorithmic fairness.
Michael Y. Li @michaelyli_
821 Followers 433 Following CS PhD @StanfordAILab @StanfordNLP advised by @noahdgoodman and Emily Fox. Prev: undergrad @princeton
Korey Wilson @koreywilsontech
86 Followers 2K Following Building coding agents @cursor_ai CS @harvard Von Neumann computing + GPUs + Quantum
Louis @Louis9687221579
109 Followers 4K Following Mainline Economics | Idea page | ramblings of a schizo
David Zhou @rectifiedzhou
341 Followers 5K Following serving humans for good ai, ml, infra @imbue_ai
λux @novasarc01
22K Followers 4K Following tensor shepherd in a non-euclidean pasture | grazing on cuda cores
Arvind Iyer @Arvind_k_Iyer
674 Followers 3K Following Learn♥️& Exploring #science | Postdoctoral Fellow @UHN @MatLupien| Ph.D under @CirielloLab | Personal Views | Alumini @IIITDelhi @unil @IGIBSocial | 🇮🇳🇨🇭
Anirudh Atmakuru @aatmakuru6
68 Followers 273 Following NLP PhD student @Cornell_CS | MS CS @UMassAmherst
Walterhu2022ᅠᅠᅠ... @walterhu3
271 Followers 6K Following
Nguyễn Vũ Thanh T�... @thanhtungg0512
38 Followers 980 Following Research Resident @ FPT Software AI Center
Darsh Bhardwaj @Rndarsh
49 Followers 574 Following Is it better to be feared or respected? I say, is it too much to ask for both?
Saurabh Shah @saurabh_shah2
4K Followers 2K Following human-ing & AI-ing @humansand prev @allen_ai @Apple @Penn 🎤dabbler of things🎸 🐈⬛enjoyer of cats 🐈 and mountains🏔️he/him
MOHD AKMAL KHAN OFFIC... @MImage53990
350 Followers 6K Following GRADUATE & MASTER in MATHEMATICS! GRADUATE in PHYSICS! We want to provide accessibility of #HIGHER EDUCATION for every INDIAN BROTHERS and SISTERS!!🙏
Jieyu Zhao @jieyuzhao11
4K Followers 995 Following Assistant Prof. @CSatUSC, @USC || Postdoc @ClipUMD || PhD from @UCLANLP, @UCLA. #NLP, #ML, #TrustworthyNLP
Yizhe Zhang @YizheZhangNLP
2K Followers 646 Following Research Scientist Manager at Apple MLR | ex-researcher @ Microsoft Research, Meta AI | PhD @ Duke. Text diffusion, latent reasoning, code generation
Samineni Soumya Rani @soumya_samineni
84 Followers 601 Following PhD in Computer Science @ ASU Interested in RL LLMs and Robotics || Past: Quantiphi, Microsoft Research. M. Tech @ IISc Bangalore & B. Tech @ NIT Warangal.
Sinjoy Saha @SinjoySaha
173 Followers 819 Following SDE @AWS AgentCore • Master's in CSE @penn_state • ex-Intern @Amazon • ex-Sr. Research Eng @SiemensHealth - MI (R&D)
Zhengyang Qi @qi_zhengyang
1K Followers 6K Following Research @SnorkelAI | Previously: @Scale_AI Multiturn RL, reward modeling, interactive environments, artificial social intelligence
Aniketh @aniketthh
123 Followers 656 Following pre-doc researcher @tcsresearch, prev: cs+design @iiitdelhi
Institute for Foundat... @MLFoundations
1K Followers 2K Following NSF AI Institute with researchers from @UTAustin, @UW, @WichitaState, @MSFTResearch, @UCBerkeley, @UCLA, @sfiscience, @Stanford, @Caltech, @ASU
Yuval Kansal @yuvalkansal
335 Followers 332 Following PhDing @Princeton. Fiddling with agents @AiatMeta. Interested in reliability and reasoning in models. Previously @UNSW Sydney
Alfonso Amayuelas @AlfonAmayuelas
986 Followers 1K Following Building an AI society 🦾 Stanford Digital Economy Lab @DigEconLab & @idai_institute CS PhD Student at @UCSB_AI @ucsantabarbara AI/NLP 😎🌊🏄🏻♂️
Praveen Kumar @imPraveenk
2K Followers 8K Following Building & Shipping 🌏 | AI & ML | Responsible AI | Physics for Entertainment | Innovation for Fixing | Nature for Learning | IISc, BLR; NIT, JSR 🌿
Tengxiao Liu @TengxiaoLiu
247 Followers 546 Following PhD student @ucsbNLP | Previous intern @Google @aws
Alyssa Unell @AlyssaUnell
348 Followers 207 Following CS PhD Student @StanfordAILab. Previously @MIT_CSAIL. Evals and reliability of AI for healthcare.
Bhagyesh Kumar @invi_bhagyesh
675 Followers 7K Following "it should be raining" maths ∩ AISafety, sophomore, I play dice, working w/ @lionellevine, currently on exchange @USTC
Robert Scoble @Scobleizer
600K Followers 57K Following San Francisco/Silicon Valley AI | Robots, holodecks, BCIs, analysis of new things | Ex-Microsoft, Rackspace, Fast Company | Wrote eight books about the future.
Yann LeCun @ylecun
1.3M Followers 793 Following Founder/Chair, AMI Labs; Professor, NYU; Partner, 224 Ventures; Ex-Chief AI Scientist, Meta. Researcher in AI, ML, Robotics, etc. ACM Turing Award Laureate.
(((ل()(ل() 'yoav)))... @yoavgo
87K Followers 2K Following
William Wang @WilliamWangNLP
22K Followers 771 Following CEO & Founder, @AlphaDesignAI. We make https://t.co/1LfDYicsF2 I'm also Mellichamp Chair Prof. at UCSB CS. PhD @ CMU SCS.
AK @_akhaliq
519K Followers 3K Following AI research paper tweets, ML @Gradio (acq. by @HuggingFace 🤗) dm for promo ,submit papers here: https://t.co/UzmYN5XOCi
Ai2 @allen_ai
86K Followers 447 Following Breakthrough AI to solve the world's biggest problems. › Join us: https://t.co/MjUpZpKPXJ › Newsletter: https://t.co/k9gGznstwj
Yi Tay @YiTayML
59K Followers 87 Following research scientist @googledeepmind ✨♊, model co-lead/captain of gemini deepthink imo gold medal 🥇, opinions are my own.
Yoav Artzi @yoavartzi
20K Followers 190 Following Research/prof @cs_cornell + @cornell_tech🚡 / https://t.co/9YnWry86x0 / researcher @GoogleDeepMind / building @COLM_conf / ex @arxiv
Percy Liang @percyliang
117K Followers 425 Following professor of computer science @Stanford @stanfordnlp, co-founder of @togethercompute, https://t.co/7R5THVnJ6u, @simile_ai, pianist
Riley Goodside @goodside
224K Followers 4K Following Mostly screenshots of chatbots since 2022. Formerly: Google DeepMind, Scale.
Kyunghyun Cho @kchonyc
88K Followers 2K Following a mediocre combination of a mediocre scientist and a mediocre advisor at @nyuniversity (@CILVRatNYU)
rishi @RishiBommasani
8K Followers 3K Following Senior Research Scholar @StanfordHAI focused on economics and governance of frontier AI Previous: Stanford CS PhD @percyliang @jurafsky, Cornell CS
François Chollet @fchollet
718K Followers 828 Following Co-founder @ndea. Co-founder @arcprize. Creator of Keras and ARC-AGI. Author of 'Deep Learning with Python'.
Luca Soldaini 🎀 @soldni
13K Followers 1K Following data mines are my passion ⛏️ mts @MicrosoftAI / ex Olmo co-lead @allen_ai / posts mine, leave my employer alone / 🌈
Mark Dredze @mdredze
6K Followers 791 Following John C Malone Professor at @JohnsHopkins @JHUCompSci @jhuclsp @jhumceh; Part time @techatbloomberg (tweets my own) Director @hopkinsdsai @mdredze.bsky.social🦋
Graham Neubig @gneubig
46K Followers 787 Following Associate professor @LTIatCMU. Co-founder/chief scientist @OpenHandsDev. I mostly work on modeling language.
Pan Lu @lupantech
7K Followers 1K Following Postdoc @Stanford @stanfordnlp | PhD @CS_UCLA @uclanlp Amazon/Bloomberg/Qualcomm Fellows | Ex @Tsinghua_Uni @Microsoft | Eubiota/AgentFlow/MathVista/ScienceQA
Sudip Roy @sudip_r0y
2K Followers 172 Following Building continuously learning intelligence @adaption_ai | ex- @GoogleDeepMind & @Cohere | Inference, efficient AI, ML for systems ↔ systems for ML
jietang @jietang
69K Followers 413 Following Professor @ Tsinghua, Founder of https://t.co/3IaQ4CI5W3. AGI, LLM. “The value of a man should be seen in what he gives and not in what he is able to receive.”―Einstein
Siddarth @siddarthv66
2K Followers 926 Following Post-training @MistralAI, PhD @Mila_Quebec | RLing LLMs
Discovery Loop @DiscoLoopAI
54K Followers 4 Following Automating discovery to accelerate science and engineering for the world.
Sanjay Ghemawat @Sanjay_Ghemawat
41K Followers 8 Following
Infini-AI-Lab @InfiniAILab
2K Followers 40 Following
Yazhe Li @yazhe_li
445 Followers 170 Following MTS @Microsoft AI. ex: DeepMind; Gatsby Computational Neuroscience Unit
elie @eliebakouch
23K Followers 4K Following training llm @PrimeIntellect (prev: @huggingface) anon feedback: https://t.co/JmMh7Sg3mL
Lee Robinson @leerob
281K Followers 872 Following Model behavior @SpaceXAI. Helping train useful models.
Binfeng Xu @billxbf
1K Followers 233 Following AI @nvidia | solving agentic coding, computer-use and continual learning with RL. Views are mine.
The OpenAI Foundation @FoundationOAI
9K Followers 0 Following OpenAI was founded in 2015 as a nonprofit; its mission is to ensure artificial general intelligence benefits all of humanity.
Zexuan Zhong @ZexuanZhong
3K Followers 758 Following Post-training & RL MSL TBD lab @Meta | prev @xAI @PrincetonCS
Recursive @Recursive_SI
8K Followers 0 Following Recursive self-improving superintelligence to automate knowledge discovery.
Mehrdad Farajtabar @MFarajtabar
10K Followers 244 Following Research Scientist at @Apple, prev @DeepMind, prev @GeorgiaTech
Apurva Gandhi @apurvasgandhi
680 Followers 168 Following PhD Student @SCSatCMU w/ @aviral_kumar2 & @gneubig | RL @NVIDIA Research. Prev: Senior ML Scientist @microsoft. Webpage: https://t.co/7fA8slEcxm
Yuchen He @YuchenHe07
4K Followers 936 Following learning @humansand & prev xai,oai,meta,apple,uiuc,utaustin
Andi Peng @TheAndiPenguin
8K Followers 995 Following prev. &i@humans& (co-founder), building claude @AnthropicAI
Reiner Pope @reinerpope
19K Followers 458 Following CEO and founder, @MatXComputing, developing high throughput chips tailored for LLMs
Cody Blakeney @code_star
8K Followers 2K Following Leading research at @arcee_ai | Formerly Data Research Lead @DbrxMosaicAI | Visiting Researcher @meta | Ph.D | #TXSTFOOTBALL fan | https://t.co/4G6Jf3b0V4
Arcee.ai @arcee_ai
15K Followers 449 Following Open models from an American research lab. https://t.co/6gcSwU5nk3
Ion Stoica @istoica05
9K Followers 20 Following Professor at UC Berkeley, co-founder of Databricks, Anyscale, LMArena, Conviva.
Timothy Lillicrap @countzerozzz
1K Followers 78 Following Artificial Intelliegence & Neuroscience Research Scientist @ DeepMind & UCL
Perplexity @perplexity_ai
501K Followers 78 Following Curiosity changes everything. Download our free app on iOS, Mac, Windows, and Android.
Michael Y. Li @michaelyli_
821 Followers 433 Following CS PhD @StanfordAILab @StanfordNLP advised by @noahdgoodman and Emily Fox. Prev: undergrad @princeton
Yinghui He @yinghui_he_
2K Followers 1K Following Princeton Computer Science PhD @PrincetonPLI. Research Intern @NVIDIA. Previously CS @UMich @sjtu1896
Xin Eric Wang @xwang_lk
22K Followers 1K Following Professor @ucsantabarbara. Head of Research @SimularAI. Director @ucsbcrml @UCSB_AI. Build the Science of Multimodal AI Agents. AI for Humanity in the long run.
Rohan Pandey @khoomeik
43K Followers 3K Following descending cross-entropy to ascend entropy @PeriodicLabs || prev research @OpenAI @CarnegieMellon '23
ClaudeDevs @ClaudeDevs
663K Followers 2 Following Official updates for developers building with @ClaudeAI
Nous Research @NousResearch
258K Followers 28 Following A bunch of nerds making progress https://t.co/vrD0aDIGDQ
Trapit Bansal @TrapitBansal
31K Followers 265 Following AI Research @Meta, Founding Member TBD Lab | Previously @OpenAI, co-creator of OpenAI o-series models (thinking in LLMs)
Jiahui Yu @jiahuiyu
28K Followers 1K Following Lead Multimodal @Meta Superintelligence Lab. Previously led Perception @OpenAI; co-led Gemini Multimodal @GoogleDeepMind.
Shengjia Zhao @shengjia_zhao
53K Followers 234 Following Chief Scientist @ Meta MSL. Formerly MTS @ OpenAI, PhD @ Stanford. I train models. All opinions my own.
Hongyu Ren @ren_hongyu
23K Followers 725 Following research @meta superintelligence. CS PhD @stanford. prev @openai.
Alexandr Wang @alexandr_wang
633K Followers 880 Following chief ai officer @meta, founder meta superintelligence labs, founder @scale_ai. rational in the fullness of time
Joseph Suarez 🐡 @jsuarez
31K Followers 124 Following I build sane open-source RL tools. MIT PhD, creator of Neural MMO and founder of PufferAI. DM for business: non-LLM sim engineering, RL R&D, infra & support.
Ronan Farrow @RonanFarrow
971K Followers 5K Following Pulitzer-winning investigative journo @newyorker. Documentaries @hbo. Ex-diplomat. Bad lawyer. Disused PhD. Tips: [email protected].
Nika Haghtalab @nhaghtal
5K Followers 306 Following Associate Professor @Berkeley_EECS. NeurIPS 2026 Program co-Chair. Research on Foundations of ML and AI. Sloan and Schmidt AI2050 fellow.
Jing Yu Koh @kohjingyu
7K Followers 397 Following ML PhD @CarnegieMellon. Previously: computer use and multimodal at MSL TBD Lab, text-to-image generation @GoogleAI. Opinions my own. 🇸🇬
David Brandfonbrener @brandfonbrener
2K Followers 715 Following member of technical staff @AnthropicAI. previously: research scientist, FAIR @AIatMeta, fellow @KempnerInst @Harvard, phd @nyu_courant
Ching-An Cheng @chinganc_rl
2K Followers 103 Following Senior Research Scientist at @Google Research, working on usable theory and algorithms for Reinforcement Learning, Generative Optimization, and RoboticsYou might like
