@Microsoft / PhD @Brown_NLP & @health_nlp / interpretability 🔁 human mechanisms / Prev: @cohere / she/they
Joined October 2014
- Tweets611
- Following1.7K
- Followers1K
- Likes4.9K
📢Our summer intern apps close in 2 days!
We're looking for interns interested in long-context modeling, retrieval, and agentic systems to explore exciting research problems at enterprise scale.
Apply here 👇
apply.careers.microsoft.com/…
(Full-time return opportunities available)
Ruochen Zhang retweeted
We let agent swarms collaborate for 10K hours in simulated agent-owned GitHubs.
Gemini-3.8-flash collaborate the best while gpt-6-astra scores the worst among frontier models tested⁉️
What are the elements of successful agent collaboration at scale? Check World of SWECraft 🧵
I wish I had the chance to work with these people if I do it all over again :))
The UT comp ling faculty (@jessyjli, @kanishkamisra, @kmahowald & myself) are looking for PhD students to join us! sites.utexas.edu/compling/ 🤘
We are an interdisciplinary group with interests across linguistics, cogsci and NLP. Apply by 12/1: liberalarts.utexas.edu/lingu…
Check out Zoey’s new work on figuring out why LLMs have language preferences! It’s been so amazing working on this together during the summer 😎
If you are working on long-context/retrieval/agentic enterprise scenarios?
Would love to chat and apply to intern with us here 🤗 (apply.careers.microsoft.com/…)
(there is also postdoc opening on the team!)
Going to COLM next week after missing so many fun conferences this year 🥹pls hmu if you are in SF 🌉
(also potentially scouting for interns 🤫)
Ruochen Zhang retweeted
I'm on the faculty job market for Fall 2027 assistant professorship. I work on the science of AI/NLP evaluation, which is currently undergoing a crisis:
Ruochen Zhang retweeted
We share our early research on building Software World - a "GitHub" run by agents. We deploy agents for Python packages in a dependency chain, and each agent is tasked with collaborating with others and optimizing the package it owns.
Ruochen Zhang retweeted
1/ How does a vision-language model know which part of its input is image and which part is text?
In our #COLM2026 paper, we study source-modality monitoring: how VLMs track where information comes from, and how they bind words like “image” or “text” to the right source.
🧵👇
Ruochen Zhang retweeted
We tried to audit how people are actually using AI assistants beyond what company reports show. Check it out here!
📍ai-observatory.org
1/ How are people really using AI?
Today, researchers across @MIT, @Stanford, + 12 academic institutions are launching the Public AI Observatory: public infrastructure for measuring how people actually use AI assistants in the wild.
🔗 ai-observatory.org/
📜 dataprovenance.org/ai_observ…
Ruochen Zhang retweeted
I will be presenting this paper at ICML mech interp workshop in-person next Friday, July 10th! (11:00-12:00 am, poster board 214). Come or dm me to chat about representational geometry and predicting model behaviors!
We don’t always know what problems are hard for LLMs. So devs evaluate on tasks HUMANS find hard or on broad benchmarks. What if we could instead anticipate which scenarios a model will fail on—all without evaluating specific input examples?
🧵NEW PAPER by @jenniferlumeng &al
Ruochen Zhang retweeted
So we're building a hard, multilingual multimodal browsing agent benchmark. Data is mostly done, just benchmarking models now. The questions are diabolical😈
Paper & data release coming soon! hyper-browsecomp.github.io
At ICML right now, hit me up if you are interested!
also...
Ruochen Zhang retweeted
Come join @lltjuatja, @kmahowald, @jessyjli, and me at UT Linguistics! Lindia and I will be around at ACL in SD, and Kyle and Jessy will be with us in spirit!🤘
My lab: kanishka.website/lab/
I will be at #ACL2026 from July 2-7! Along with checking out the conference and hanging around SD, I'm also ✨recruiting students for Fall 2027 @ UT Austin Linguistics (along with @kanishkamisra)✨
Some topics I'm really excited about below ⬇️
Ruochen Zhang retweeted
I will be at #ACL2026 from July 2-7! Along with checking out the conference and hanging around SD, I'm also ✨recruiting students for Fall 2027 @ UT Austin Linguistics (along with @kanishkamisra)✨
Some topics I'm really excited about below ⬇️
Ruochen Zhang retweeted
Hiring Research Interns | Multilingual Agents @ Qwen Team
I am seeking Research Interns to work on multilingual agent research in Beijing.
Research topics include:
- Multilingual large language models and agent systems
- Cross-lingual reasoning, planning, and tool use
- Post-training, alignment, and synthetic data generation for multilingual settings
- Evaluation of multilingual agents in complex, real-world tasks
Qualifications:
- Currently enrolled in a PhD program in Computer Science, NLP, AI, or a related technical field
- Strong research background in LLMs, NLP, agents, machine learning, or multilingual AI
- Experience conducting independent research, including problem formulation, experimentation, and empirical analysis
- Familiarity with Python, PyTorch, and modern ML research workflows
- Publication record is preferred
- Fluency in Chinese is required
- Able to work onsite in Beijing
This internship offers the opportunity to contribute to cutting-edge research on multilingual foundation models and agent systems with real-world impact.
If interested, please DM me or send your resume to minghao[dot]wu[at]alibaba-inc[dot]com.
Ruochen Zhang retweeted
Benchmarks can be superficial, but model explanations and evaluations are fundamentally intertwined. What if we used interpretability as principled, scientific evaluation? If it met scientific standards?
arxiv.org/abs/2605.05508
coming to @evaluatingevals at ACL as oral 🧵
1/6
ALT A list of researchers and their affiliations, discussing the theme "Rigorous Interpretation Is a Form of Evaluation."