greenai.guide

Accueil › Annuaires › Offres de thèse en IA verte

Annuaire

Offres de thèse en IA verte

Des contrats doctoraux en IA verte, en apprentissage automatique efficace, en informatique durable et en mathématiques de l’IA efficace.

À propos de cet annuaire

18 contrats doctoraux et postes de doctorant sur l’IA verte et efficace, rassemblés à partir de sites d’universités et de recrutement. Les dates limites et les financements changent : vérifiez toujours l’annonce avant de postuler.

Ouvert

PhD Stipend in Energy-Efficient Agentic AI: Measuring When AI Is Worth Its Energy

Aalborg University, Department of Computer Science · Denmark

Measures the energy use of agentic AI systems and compares it with conventional deterministic software, to work out when an AI agent is worth the energy it consumes. Start date 1 March 2027.

Fully funded, three years; Independent Research Fund Denmark (DFF), thematic AI programme · Deadline 2026-12-05 · ['Michele Albano', 'Roman Jurowetzki (co-supervisor)']

Lire l’annonce →
Ouvert

Doctoral Researcher in Adaptive Grid-Interactive Edge Datacenter Fleets (MSCA Doctoral Network AEGIS)

Aalto University, School of Electrical Engineering · Finland

Builds end-to-end energy models for distributed AI services, from radio and transport networks to datacentre hardware, and exposes them through a fast query interface. The call may close early if filled.

Funded, three-year contract; EU Marie Skłodowska-Curie Doctoral Network AEGIS (Horizon Europe) · Deadline 2026-12-15 · ['Gopika Premsankar', 'Riku Jäntti (advisor)']

Lire l’annonce →
Ouvert

Research Studentship in Control Engineering: Sustainable and Circular AI Systems

University of Oxford, Department of Engineering Science · United Kingdom

Investigates how the rapid growth of AI can be supported by more sustainable and circular hardware and infrastructure, looking beyond operational energy to materials and lifecycle impacts. Starts October 2027.

Funded 3.5-year DPhil studentship, stipend £21,805 p.a.; open to UK, EU and international students · Deadline 2026-12-02 · ['Lucia Corsini']

Lire l’annonce →
Ouvert

High-Efficiency Power Electronics for AI Data Centres and Net-Zero Grids

University of Warwick · United Kingdom

Develops medium-voltage solid-state transformer cells to cut the energy lost in power conversion inside hyperscale AI data centres. Focuses on compact high-frequency magnetics and reliability under electrical-thermal stress.

Funded studentship for UK students, £21,805 stipend · Deadline 2026-12-01 · ['Jiaqi Yan', 'Layi Alatise']

Lire l’annonce →
Ouvert

Uncertainty-Aware Synchronisation for Sustainable Data Centre Digital Twins

University of Exeter · United Kingdom

Treats digital-twin synchronisation for renewable-powered data centres as a decision problem under forecast uncertainty, so twins update only when needed. Aims to save computing resources while keeping the twin faithful to the real facility.

Funded for UK students: fees plus stipend of £21,805 p.a. (2026/27) · Deadline 2026-12-08

Lire l’annonce →
Ouvert

R2T2-Med: Resource-Aware AI for Scalable and Deployable Medical Image Analysis

University of Exeter · United Kingdom

Builds a train-once model family for 2D and 3D medical imaging that adapts its compute to case complexity and available hardware. Targets wasted computation in current models without losing clinically important detail.

Funded for UK and international students: fees plus stipend of £21,805 p.a. (2026/27) · Deadline 2026-12-08

Lire l’annonce →
Ouvert

Tiny Multimodal Learning Under Resource Constraints

Manchester Metropolitan University · United Kingdom

Asks how multimodal AI (audio, motion, images, sensors) can work with minimal, efficient representations when computation, memory or energy are limited. Studies the trade-offs between accuracy, efficiency and robustness.

Funded three-year studentship; home and international fees covered plus standard UKRI-rate stipend · Deadline 2027-02-12

Lire l’annonce →
Ouvert

Neuromorphic Systems for Low-latency and Efficient Multimodal Processing (FSE President's Doctoral Scholarship)

The University of Manchester · United Kingdom

Develops brain-inspired neuromorphic algorithms for efficient multimodal processing, aiming at intelligent processing on a power budget close to the brain's. Open to home and international applicants.

Four-year scholarship at UKRI rate plus £1,000 p.a. and up to £20,000 research support; project competes for funding · Deadline 2026-12-18 · ['Luca Peres']

Lire l’annonce →
Ouvert

Efficient Physical AI for Real-World Robotics

Durham University (with Intel) · United Kingdom

Develops capable robot AI (foundation and vision-language-action models, planning, data-efficient learning) that runs within tight limits on compute, memory, latency and power. Moves away from dependence on very large models and cloud inference.

Fully funded by Intel: home tuition fees and tax-free stipend; Home fee-status applicants only · Deadline 2027-01-19 · ['Amir Atapour-Abarghouei', 'Toby Breckon', 'Samet Akcay (Intel)']

Lire l’annonce →
Ouvert

Terahertz Magnetic Recording for Green Data Storage Technology

Lancaster University · United Kingdom

Physics PhD working towards ultra-low-energy, ultrafast memory switched by terahertz pulses, to tackle the energy and heat burden of data-centre storage. Admission October 2027; needs a Master's in Physics.

Fully funded for four years for UK, EU and international students (Green Future Fellowship project) · Deadline 2027-01-31 · ['Rostislav Mikhaylovskiy']

Lire l’annonce →
Ouvert

Machine Learning the Building Blocks of the Weather Forecast

University of Exeter (with the Met Office) · United Kingdom

Breaks the weather forecast into its component processes and matches a machine-learning architecture to each, aiming for models that are more efficient, explainable and robust under a warming climate.

Funded for eligible students: home tuition fees plus tax-free stipend · Deadline 2027-01-07

Lire l’annonce →
Ouvert

Improve the Energy Sustainability of AI Data Centers in Future Energy System

University of Edinburgh, School of Engineering · United Kingdom

Develops real-time scheduling of AI computing jobs to cut data-centre carbon footprints, and explores how data centres can offer grid services such as demand response. Open until filled.

Self-funded, or applicants seeking University of Edinburgh or external scholarships · Deadline 2026-12-31 · ['Wei Sun', 'Gareth Harrison']

Lire l’annonce →
Ouvert

Intelligent Earth: UKRI AI Centre for Doctoral Training in AI for the Environment (2027 entry)

University of Oxford · United Kingdom

Four-year doctoral training programme applying AI to environmental and Earth-system challenges, with cohort-based training. Deadline is midday UK time; start September 2027.

UKRI AI CDT; applicants are automatically considered for funded places · Deadline 2027-01-06

Lire l’annonce →
Ouvert

Testing the Ability of AI-Based Regional Models to Capture Hydroclimate Changes of the Geologic Past (UNRISK CDT)

University College London, Department of Geography (NERC CDT UNRISK) · United Kingdom

Tests whether cheap AI-based climate downscaling can reproduce past hydroclimate shifts seen in costly high-resolution simulations, and whether its uncertainty estimates hold up. The CDT invites supervisor contact from October to late December.

NERC Centre for Doctoral Training UNRISK studentship (competitive; 18 funded places for 2027 entry) · ['Chris Brierley']

Lire l’annonce →
Vérifier le statut

PhD Scholarship in Green Computing

RMIT University · Australia

Researches ways to cut the energy use, carbon footprint and e-waste of AI-driven digital systems across cloud, edge and 5G/6G infrastructure. The page states eligibility ends 30 December 2026 but gives no closing date.

Funded: AUD 36,245 p.a.; two scholarships; domestic and international applicants

Lire l’annonce →
Vérifier le statut

The Hidden Cost of AI Agents: Towards Energy-Efficient Agentic Systems

Curtin University · Australia

Builds a framework to make multi-step LLM agents more energy-efficient through smarter planning, selective use of small and large models and fewer redundant calls. Entry is by expression of interest.

No scholarship attached; international applicants self-fund; domestic students may get an RTP fee offset · ['Sajib Mistry (contact)']

Lire l’annonce →
Vérifier le statut

Energy-Efficient Architectures for Embedded AI in Space

Politecnico di Torino, Image Processing Lab · Italy

Develops low-precision neural networks (e.g. ternary networks) that approach state-of-the-art accuracy at a fraction of the energy, for onboard satellite computing. Selection was targeted for April 2026 with a November 2026 start.

Funded through the ERC Advanced Grant IntelliSwarm, with a top-up above the standard PoliTo PhD salary · ['Enrico Magli']

Lire l’annonce →
Vérifier le statut

PhD Position: Sustainable Generative AI for Digital Sovereignty

Delft University of Technology, Faculty of Technology, Policy & Management · Netherlands

Studies how to run generative AI sustainably and under European control, evaluating foundation models and designing efficient inference across the computing continuum. Recently closed; listed as an example of the area.

Funded four-year position within an EU Horizon project; salary per Dutch universities CAO · Deadline 2026-11-01 · ['Aaron Ding', 'Marcela Tuler de Oliveira', 'Marijn Janssen']

Lire l’annonce →

Questions fréquentes

Publier une offre de thèse est-il payant ?

Non. Les universités, les écoles doctorales et les directeurs de thèse peuvent publier une offre de thèse gratuitement.

Ces fiches sont-elles à jour ?

Chaque offre a été vérifiée sur la page de l’établissement ou du recruteur à la date indiquée dans nos données ouvertes. Les offres marquées « Vérifier le statut » sont peut-être closes.

Rejoignez-nous

Aidez à réduire l’impact de l’IA sur la planète

Recevez par e-mail, environ une fois par mois, les nouveaux guides, appels à financement et fiches des annuaires. Ou ajoutez votre produit, cours, thèse ou recherche aux annuaires.

Aucun suivi. Désinscription en un clic. Autres façons de participer · Confidentialité