Open tracks
PhD (UofT, ELLIS, MPI) + Postdoc
Application review
Rolling, monthly batches
Locations
Toronto · Tübingen · Zürich
CS PhD Recruitment Roundup — Week of May 21–28, 2026
Five labs actively recruiting PhD students or postdocs this week: Zhijing Jin's causal-LLM lab (UofT/ETH/MPI), Mohit Bansal's MURGe-Lab (UNC Chapel Hill), Haohan Wang's trustworthy-ML lab (UIUC), Allen AI Predoctoral YI positions, and Zhiyong Cui's AI+transportation lab (Beihang). Research directions and lab culture notes included.
Five labs with publicly posted openings this week, across NLP, trustworthy ML, multimodal agents, privacy, and AI+transportation. Research directions and what each lab is actually like to work in.
1. Jinesis Lab — causal LLMs, AI safety, multi-agent systems
PI: Zhijing Jin (Assistant Professor, University of Toronto; CIFAR AI Chair; affiliated with MPI Tübingen and ETH Zürich)
Conference track record: NeurIPS, ACL, EMNLP (causal reasoning, social NLP, AI safety)
What they work on: The lab splits its attention across three areas: applying formal causal inference to LLMs, studying multi-agent LLM behavior, and AI safety for societal benefit. The framing is deliberately interdisciplinary — Jin recruits people who can think across philosophy, statistics, or political science alongside ML. Research runs from theory-facing (causal identifiability in language models) to applied (democratic defense against LLM misuse).
Lab culture notes: Jin runs a distributed team across Toronto, Tübingen, and Zürich; students can co-supervise with Bernhard Schölkopf at MPI. The openings page explicitly welcomes applicants from regions where top-conference infrastructure is sparse (Jin treats pre-PhD research collaborations as a fairer evaluation signal than publication counts). Applications from Iran, Russia, and China require navigating Canadian and German visa policy on sensitive technology — the lab flags this directly.
Open positions:
- PhD — University of Toronto (CS, ECE, or Stats; rolling applications, pre-PhD form reviewed monthly, last batch: April 29)
- PhD — ELLIS program (co-supervised with Schölkopf; apply via ELLIS PhD 2025 call)
- PhD — Max Planck Institute (via IMPRS-IS or ELLIS channels, CC Jin on application)
- Postdoc — open but low demand currently; inquire by email with subject
[Postdoc interest match]
Salary / stipend: Not disclosed on openings page
Location: Toronto, Canada / Tübingen, Germany / Zürich, Switzerland (distributed)
Deadline: Rolling (pre-PhD form reviewed monthly)
Apply / contact: https://zhijing-jin.com/home/index.php/openings/
Loading stats card…
2. MURGe-Lab (Bansal Lab) — multimodal NLP, agents, world models
PI: Mohit Bansal (Parker Distinguished Professor, UNC Chapel Hill; ACL Fellow; AAAI Fellow; PECASE recipient)
Conference track record: ACL, EMNLP, NeurIPS, CVPR — long-horizon reasoning, visual grounding, video generation, multi-agent communication
What they work on: The lab covers a wide range of multimodal and language tasks: controllable video generation, long-video understanding, embodied agents with skill abstraction, reward modeling, model interpretability and editing, and unified multimodal generation. Recent directions include world models and "any-to-any" generation.
Lab culture notes: Bansal runs one of the larger NLP/multimodal groups in the US. The lab is part of the ENGAGE NSF-AI Institute, which means students get exposure to a broader research ecosystem and funded collaborations. The prospective students page notes the team "ranks at the top in NLP" and emphasizes collaboration opportunities. Bansal explicitly encourages diverse applicants and offers fee waivers.
Open positions:
- PhD — UNC Chapel Hill CS (Fall 2026 cycle; apply via official CS admissions, mark NLP as area and name Bansal in personal statement)
- Postdoc — 2026–2027 (posted; contact Bansal directly and see PDF advertisement on lab site)
Salary / stipend: Not disclosed; standard UNC RA/TA packages apply
Location: Chapel Hill, NC, USA (on-site)
Deadline: PhD applications: rolling through March 12 for RA/TA; December 10 for fellowship consideration. Postdoc: open now
Apply / contact: https://www.cs.unc.edu/~mbansal/ — see "Prospective Students" and postdoc PDF link
Loading stats card…
3. Haohan Wang Lab — agentic AI, trustworthy ML, computational biology
PI: Haohan Wang (Assistant Professor, School of Information Sciences, UIUC)
Conference track record: NeurIPS, ICML (trustworthy ML, spurious feature analysis, robust learning)
What they work on: Three overlapping threads: building agentic AI architectures that can operate autonomously on complex scientific data; studying safety and security of those agents (adversarial misuse, jailbreaking); and applying trustworthy ML to computational biology — particularly genetics of Alzheimer's and cancer. The lab is also developing MyDataPilot, a free, low-barrier software tool for domain experts.
Lab culture notes: Wang explicitly says he is "always happy to chat" with interested students, whether about formal admission or remote research collaborations. The lab is newer (Wang joined UIUC after completing his PhD under Eric Xing at CMU), which means students get close mentorship in a small-group environment. An open collaborative model with other global trustworthy ML researchers is explicitly part of the lab's identity.
Open positions:
- PhD — UIUC iSchool (direct contact welcome; also recruiting for MyDataPilot development)
- Research collaboration (remote; for students building toward a PhD application)
Salary / stipend: Not disclosed
Location: Urbana-Champaign, IL, USA
Deadline: Not stated; rolling contact
Apply / contact: https://www.cs.cmu.edu/~haohanw/ — email with research interests and CV
4. Allen Institute for AI (Ai2) — Predoctoral Young Investigators
Organization: Allen Institute for AI (Ai2) — a nonprofit AI research institute founded by Paul Allen. Known for Semantic Scholar, OLMo, AllenNLP, and Macaw; strong publication record at ACL, NeurIPS, ICLR, CVPR.
What the role involves: Ai2's Predoctoral Young Investigator (YI) program places early-career researchers alongside senior scientists on frontier projects, with the aim of co-authoring papers at top venues before PhD admission. Predoctoral YIs are full employees, not interns.
Lab culture notes: Ai2 runs a collaborative, flat-ish structure that is unusual for a research organization of its size (~150 staff scientists and YIs). The program is explicitly designed as a bridge for candidates who want research experience before committing to a PhD program, and Ai2 has a track record of YIs going on to PhD programs at top CS departments. International applicants are welcome; Ai2 sponsors visas (processing typically 2–3 months post-offer).
Open positions:
- Predoctoral Young Investigator — Open Ecosystem (Seattle, WA)
- Predoctoral Young Investigator — Robotics (Seattle, WA)
Salary / stipend: Competitive full employee compensation (Ai2 does not publish a salary band)
Location: Seattle, WA, USA (on-site)
Deadline: Applications reviewed on a rolling basis; formal applications only (no email inquiries)
Apply / contact: https://allenai.org/careers — search "Predoctoral" to find both roles
Loading stats card…
5. Zhiyong Cui Lab — AI + transportation, autonomous driving
PI: Zhiyong Cui (Professor, School of Transportation Science and Engineering, Beihang University)
Conference track record: NeurIPS, AAAI, IEEE transactions (spatial-temporal learning, graph neural networks for traffic, autonomous driving)
What they work on: Four overlapping areas at the intersection of ML and transportation: (1) foundation models for transportation using LLMs and multimodal-language models; (2) autonomous driving — perception, prediction, planning, and scenario generation; (3) spatial-temporal intelligence for urban computing and traffic forecasting; (4) multi-agent learning for autonomous vehicles, traffic signal control, UAVs, and robotics.
Lab culture notes: Cui's lab is one of the more active Chinese university groups bridging deep learning and transportation engineering. The group publishes regularly at ML venues (NeurIPS, AAAI) as well as transportation-specific journals. Beihang has a strong aerospace and systems engineering culture, which means the lab's work tends toward deployable, safety-critical applications rather than purely benchmark-driven research.
Open positions:
- PhD — Beihang University (1–2 positions; admit year: 2025 or 2026)
Salary / stipend: Standard Chinese graduate stipend (not disclosed)
Location: Beijing, China
Deadline: Not stated; contact directly
Apply / contact: https://zhiyongc.github.io/ — email Zhiyong Cui with research interests
How to read this roundup
Each entry above was sourced from the lab's own website or openings page, visited this week. Deadlines and availability can change; check the linked URL before applying.
This channel covers PhD and postdoc openings posted by authors who have published at NeurIPS, ICML, ICLR, CVPR, or ACL. It runs weekly on Mondays covering the prior seven days.
Add more perspectives or context around this Drop.