CS Top-Conference PhD Recruitment Roundup — Week of May 15, 2026

This week's roundup surfaces 11 confirmed PhD and postdoc openings from 9 advisors who have all published at NeurIPS, ICML, ICLR, CVPR, or ACL between 2022 and 2025. All positions are at European institutions. Three deadlines close within 5 days of publication (May 20: TU Delft causal ML, Aarhus Edge TTA; May 21: Copenhagen BioML). Positions are grouped by research direction — NLP/LLM security, ML for biology/biomedicine, computer vision/adaptive AI, and others.

This week's scan confirmed 11 confirmed openings across 9 advisors, all at European institutions. Several listings cover 2 seats, so the total number of available PhD or postdoc slots is higher than 11. Every PI listed here has published at NeurIPS, ICML, ICLR, CVPR, or ACL between 2022 and 2025. Deadlines cluster between May 20 and June 8 — three positions close within the next five days.

One note on scope: the Twitter/X discovery unit timed out during research, so direct PI posts on that platform are not reflected this week. All confirmed positions were sourced from official job portals, lab websites, and DBLP publication records. No US university positions meeting the top-venue criteria were found — PhD admissions at US institutions are typically handled through centralized departmental portals rather than individual PI postings.

Advisor tier key used throughout this roundup:

⭐⭐⭐ — 5+ qualifying papers at target venues (2022–2025)
⭐⭐ — 3–4 qualifying papers
⭐ — 1–2 qualifying papers

NLP, LLM security, and interpretability

PhD — LLM security & mechanistic interpretability | CopeNLU, University of Copenhagen ⭐⭐⭐

Advisor: Isabelle Augenstein (Full Professor, Deputy Head of Research, DIKU — the Department of Computer Science at the University of Copenhagen) — 12 papers at ACL/NeurIPS 2022–2025; 10,900+ citations 1
Position type: PhD (fully funded, 3 years)
Research direction: Mechanistic interpretability methods for curbing false information attacks on LLMs; LLM security, explainable AI (XAI)
Project: "A Mechanistic Framework for Mitigating the Susceptibility of LLMs to Learning False Information" 2
Funding: Independent Research Foundation Denmark (2.96 million DKK); collaboration opportunity with NVIDIA as academic partner
Co-supervisor: Pepa Atanasova
Requirements: Master's degree in CS or a related field; ML/NLP background preferred
Start date: September 2026 (or as soon as possible)
Deadline: May 31, 2026
Apply: employment.ku.dk

Lab notes: CopeNLU (Copenhagen Natural Language Understanding) is one of Europe's strongest NLP groups, with multiple active PhD students, postdocs, and international collaborators. The group publishes consistently at ACL, EMNLP, and NeurIPS, with a focus on responsible, accountable NLP. The lab's website includes a detailed "Why UCPH" page covering working conditions, funding, and career paths. 2

Postdoc — LLM security & mechanistic interpretability | CopeNLU, University of Copenhagen ⭐⭐⭐

Advisor: Isabelle Augenstein (same as above)
Position type: Postdoc (fully funded, 3 years)
Research direction: Characterising the spectrum of false information encoded by LLMs; developing mechanistic mitigation methods
Project: Same as the PhD position above
Requirements: PhD in CS or a related field; background in NLP, LLM security, and/or XAI
Start date: September 2026 (or as soon as possible)
Deadline: May 31, 2026
Apply: employment.ku.dk

Lab notes: The PhD and postdoc positions are part of the same grant and will work together within CopeNLU. Both roles offer a collaboration pathway with NVIDIA. 2

PhD — applied ML & NLP for legal AI | ETH Zurich Center for Law & Economics ⭐

Advisor: Jakob Merane (Senior Research Associate, ETH Zurich Center for Law & Economics) — 1 paper at ACL 2025 ("SwiLTra-Bench: The Swiss Legal Translation Benchmark"); also published at EMNLP 2025 3
Position type: PhD
Research direction: Applied ML and NLP for the Swiss judiciary; LLM/RAG-based legal decision-support systems developed in collaboration with Swiss courts
Project: SNSF-funded (Swiss National Science Foundation) "Responsible AI for the Swiss Judiciary" 4
Requirements: MSc in CS, Data Science, or a closely related field; strong ML/NLP background and practical engineering skills; German language skills are a plus
Location: ETH Zurich Center for Law & Economics, downtown Zurich
Deadline: Not stated — apply as soon as possible 4
Apply: ETH Zurich application portal — questions only to [email protected]

Lab notes: ETH Zurich consistently ranks among the world's top universities. The project sits at the intersection of NLP, legal reasoning, and civic AI — an unusual niche for CS PhD work. No application deadline is stated on the posting. 4

Machine learning for biology and biomedicine

2× PhDs — ML & AI for biology | Kobak Lab, VIB.AI / Ghent University ⭐⭐⭐

Advisor: Dmitry Kobak (Full Professor, Ghent University; Principal Investigator, VIB Center for AI & Computational Biology) — 5 papers at ICLR/ICML/NeurIPS 2022–2025; 7,000+ citations 5
Position type: PhD (fully funded, 2 positions); postdoc positions also available
Research direction: Self-supervised learning, foundation models for biology, dimensionality reduction, single-cell genomics
Requirements: MSc in CS, AI, Applied Mathematics, Physics, Bioinformatics, Bioengineering, or similar; experience with modern ML (deep learning, probabilistic modeling, generative AI); life-science experience preferred but not required; non-traditional backgrounds explicitly welcome 6
Funding: Fully funded PhD scholarship; support for national fellowship applications (e.g., FWO — Fonds Wetenschappelijk Onderzoek, the Flemish research funding agency)
Application materials: CV, motivation letter (max 1 page, include GitHub link), BSc/MSc transcripts, 2–3 referees
Deadline: June 7, 2026
Apply: jobs.vib.be — questions to [email protected]

Lab notes: Kobak Lab is a newly established group at VIB.AI, a brand-new research center in Ghent that launched in 2026 and is actively expanding. Kobak joined as group leader in April 2026, meaning early PhD students will help shape the lab's initial research directions. Shortlisted candidates are interviewed over Zoom and invited for an on-site visit in Ghent. VIB.AI is part of VIB (Vlaams Instituut voor Biotechnologie), a leading life-sciences research institute in Belgium. 6

2× PhDs — neuro-symbolic ML for biology & drug design | BioML, University of Copenhagen ⭐⭐⭐

Advisor: Wouter Boomsma (Professor, Machine Learning section, DIKU) — 4 papers at NeurIPS/ICLR 2022–2025; 4,800+ citations 7
Position type: 2 PhD fellowships (fully funded)
Research direction:
- PhD 1: Probabilistic circuits for protein family modeling (collaboration with University of Edinburgh)
- PhD 2: Literature-aware neurosymbolic modeling of GPCR (G-protein-coupled receptor, a major drug target class) interactions (collaboration with Department of Drug Design, University of Copenhagen)
Lab: BioML group at DIKU — focuses on structural bioinformatics, protein engineering, biomolecular modeling, and Bayesian optimization for biological sequence design
Requirements: Not detailed in the posting; contact [email protected] for specifics
Start date: September 1, 2026
Deadline: May 21, 2026 ⚠️ closes in 6 days
Apply: employment.ku.dk

Lab notes: BioML sits within DIKU's Machine Learning section, which also hosts CopeNLU (Augenstein) and other strong NLP and bio-ML groups. The collaborative structure of both PhD positions — one with Edinburgh, one with the Drug Design department — means students will have access to expertise beyond the immediate lab. 8

PhD — generative modelling for biomedicine | CRUK Cambridge Institute, University of Cambridge ⭐

Advisor: Hana Aliee (Group Leader, Cancer Research UK Cambridge Institute) — 1 paper at NeurIPS 2022; 2,900+ citations 9
Position type: PhD studentship (fully funded, 4 years)
Research direction: Generative modelling (diffusion models, flow matching), causal ML, GNNs, multimodal data integration for biomedical discovery
Project: "Generative Modelling for Foundational Discovery in Biomedicine"
Stipend: £22,500/year 10
Start date: January 2027
Deadline: June 8, 2026
Apply: cam.ac.uk

Lab notes: The Aliee Lab is based at the Cancer Research UK Cambridge Institute, a world-leading cancer research center on the Cambridge Biomedical Campus. The lab's work sits directly at the intersection of frontier generative AI methods and biological mechanism discovery — a rare combination of methodological novelty and immediate translational relevance. Aliee previously built her group at Helmholtz Munich before moving to Cambridge. 10

Computer vision and adaptive AI

PhD — test-time adaptation & agentic AI | A3 Lab, Aarhus University ⭐⭐⭐

Advisor: Behzad Bozorgtabar (Associate Professor, Aarhus University; Guest Researcher, Pioneer Centre for AI) — 5 papers at CVPR/ICLR 2022–2025, plus 2 NeurIPS 2025 papers pending DBLP indexing 11
Position type: PhD (fully funded)
Research direction: Test-time adaptation (TTA) for multimodal foundation models; agentic decision-making; uncertainty estimation under real-world distribution shift
Lab: A3 Lab (Adaptive & Agentic AI) — a newly established group at Aarhus University's Department of Electrical and Computer Engineering
Start date: August 1, 2026
Deadline: June 1, 2026
Apply: Aarhus University Graduate School of Technical Sciences — also posted on Nature Careers 12

Lab notes: A3 Lab is a new group, giving early PhD students significant input into the lab's research direction. Bozorgtabar is an active recruiter and posts about lab developments on LinkedIn. He also holds a guest researcher affiliation at the Pioneer Centre for AI (Denmark's national AI center), which means access to a broader research community beyond Aarhus. 13

PhD — agentic TTA for edge intelligence | A3 Lab, Aarhus University ⭐⭐⭐

Advisors: Behzad Bozorgtabar (primary, see above) + Qi Zhang (Professor, Edge Intelligence, Aarhus University) as co-supervisor
Position type: PhD (fully funded)
Research direction: Agentic test-time adaptation for efficient and reliable deployment of foundation models on edge devices; autonomous monitoring, on-the-fly adaptation, efficiency/reliability trade-offs
Start date: August 1, 2026
Deadline: May 20, 2026 ⚠️ closes in 5 days
Apply: Aarhus University Graduate School 12

Lab notes: This position adds an edge AI systems angle to the TTA research agenda of A3 Lab, through co-supervision with Qi Zhang (whose research covers edge communications and multi-agent systems). Note that Qi Zhang's own top-venue publications are from 2019–2020, outside the 2022–2025 window, so the primary advisor credential here rests with Bozorgtabar.

PhD — computer vision for multi-agent perception | ETRO, Vrije Universiteit Brussel ⭐⭐

Advisor: Nikos Deligiannis (Professor, ETRO department, Vrije Universiteit Brussel (VUB); Principal Investigator, imec) — 3 papers at CVPR/NeurIPS/ICLR 2022–2025; 4,000+ citations; PI of ERC IONIAN project 14
Position type: PhD (fully funded, ERC-funded)
Research direction: Computer vision for cooperative multi-agent perception in autonomous driving; foundation models, world models, vision-language models (VLMs), diffusion models for collaborative scene understanding
Funding: ERC IONIAN project (reinventing multi-terminal coding for intelligent machines)
Institution notes: ETRO (Electronics and Informatics) is a well-established research department at VUB (Vrije Universiteit Brussel — a major research university in Brussels), with strong industry ties through imec, the world-leading nanoelectronics research center. 15
Deadline: June 1, 2026
Apply: academicpositions.com listing, or contact [email protected] directly

Other confirmed openings

2× PhDs — ML for digitising smell | KTH Royal Institute of Technology ⭐⭐⭐

Advisor: Danica Kragic (Professor, School of Computer Science and Communication, KTH) — 6 papers at ICLR/ICML/NeurIPS 2022–2025; 26,000+ citations; invited speaker at NeurIPS 2024; Director of the Centre for Autonomous Systems; Wallenberg Scholar 16
Position type: 2 PhD positions
Research direction:
- Position 1: Machine learning for EEG time-series data (neural signals underlying olfactory perception)
- Position 2: Computer vision and video processing for olfactory AI
Project: EU project "Digitising Smell" — applying ML and robotics methods to understand and replicate olfactory perception 17
Institution: KTH Royal Institute of Technology, Stockholm — one of Europe's leading technical universities
Deadline: May 27, 2026
Apply: KTH job portal (see the AcademicPositions listing)

Lab notes: Kragic runs one of Europe's largest and most cited robotics and AI labs. Her research has recently expanded into sensory AI beyond vision, making these positions part of a genuinely novel research frontier for the group. No publicly available independent student-experience reviews were found.

PhD — causal inference & machine learning | TU Delft ⭐

Advisor: Jesse Krijthe (Assistant Professor, Pattern Recognition & Bioinformatics group, Delft University of Technology) — 1 paper at NeurIPS 2023 (main conference track): "Detecting hidden confounding in observational data using multiple environments" 18
Co-supervisor: Nan van Geloven (medical statistics researcher, LUMC — Leiden University Medical Center)
Position type: PhD (fully funded)
Research direction: Detecting and mitigating causal assumption violations, particularly positivity violations; causal inference methods with medical applications
Consortium: Part of the "Safe Causal Inference" consortium (8 PhD positions total across Dutch institutions)
Salary: €3,059–€3,881/month 19
Deadline: May 20, 2026 ⚠️ closes in 5 days (rolling admissions — final deadline)
Apply: careers.tudelft.nl — upload CV, motivation letter, and transcripts; questions to j.[email protected]

Lab notes: TU Delft (Delft University of Technology) is one of the Netherlands' top technical universities and a member of the IDEA League, a network of leading European technical universities. Krijthe's research focuses on causal inference, semi-supervised learning, and the philosophy of statistics. The consortium structure means PhD students will interact with peers at other Dutch institutions working on related causal ML problems.

Advisor snapshot

Advisor	Institution	Research area	Top-venue papers (2022–2025)	Positions open	Deadline
Isabelle Augenstein	U Copenhagen (DIKU)	LLM security, XAI, NLP	ACL ×10, NeurIPS ×2 (12 total)	PhD + Postdoc	May 31
Wouter Boomsma	U Copenhagen (DIKU)	ML for biology, drug design	NeurIPS ×3, ICLR ×1 (4 total)	2× PhD	May 21
Danica Kragic	KTH	ML, robotics, olfactory AI	ICLR ×2, ICML ×3, NeurIPS ×1 (6 total)	2× PhD	May 27
Jesse Krijthe	TU Delft	Causal ML, statistics	NeurIPS ×1	PhD	May 20
Behzad Bozorgtabar	Aarhus University	CV, TTA, agentic AI	CVPR ×2, ICLR ×3 (5 total)	2× PhD	May 20 / June 1
Nikos Deligiannis	VUB (ETRO)	CV, vision-language models, AV	CVPR ×1, NeurIPS ×1, ICLR ×1 (3 total)	PhD	June 1
Dmitry Kobak	VIB.AI / Ghent U	ML for biology, self-supervised	ICLR ×2, ICML ×2, NeurIPS ×1 (5 total)	2× PhD	June 7
Hana Aliee	CRUK / U Cambridge	Generative models, causal ML, biomedicine	NeurIPS ×1	PhD	June 8
Jakob Merane	ETH Zurich	Applied ML/NLP, legal AI	ACL ×1	PhD	Not stated

Application notes

Act now on May 20 deadlines: Jesse Krijthe (TU Delft, causal ML) and Behzad Bozorgtabar's Edge TTA PhD (Aarhus) both close on May 20. The TU Delft position uses rolling admissions — applications reviewed as received, with May 20 as the final cut-off. The Wouter Boomsma fellowships at Copenhagen close May 21.

Geographic scope: All 11 confirmed positions are at European institutions this week (Netherlands, Denmark, Belgium, Germany/Switzerland, UK, Sweden). US PhD recruiting follows a different calendar and channel (centralized departmental applications, not individual PI announcements), and no US positions from verified top-venue PIs were found through this week's scan.

Sub-field gaps: This week's confirmed openings cluster in NLP/LLM security, computer vision/adaptive AI, and ML-for-biology. No openings were confirmed in reinforcement learning, ML theory, or multi-agent RL from PIs meeting the top-venue criteria.

Note on co-supervised positions: The Aarhus Edge TTA PhD (item 7) is co-supervised by Qi Zhang, whose own top-venue publications fall outside the 2022–2025 window. The qualifying credential for that position rests with primary advisor Bozorgtabar.

参考来源

围绕这条内容继续补充观点或上下文。

登录后可发表评论。