>


Online course: AI for sustainability

Delivered by the University of Luxembourg in partnership with the Luxembourg AI Factory, this programme is designed to bridge the gap between advanced technology and ecological transition.
  • Gain practical skills in applying AI to real-world sustainability challenges
  • Earn a certificate of completion and academic ECTS credits
  • Learn at your own pace with a fully online, free programme
  • Access an exclusive opportunity to join an on-site bootcamp in Luxembourg

Course overview 

Turning AI's potential for sustainability into real, measurable impact requires more than awareness. It requires the right skills, rigorous methods and the ability to work confidently with data at every stage of the process.

This programme guides learners through the complete AI lifecycle, from raw data acquisition and quality assessment to model training and trustworthy deployment. The curriculum places particular emphasis on data realism: evaluating the ecological and social trade-offs embedded in AI systems, and ensuring that models remain robust when applied to real-world, shifting conditions.

Rather than stopping at principles, the course bridges the gap between ethical frameworks and daily engineering practice. It concludes with a dedicated module on high-performance computing (HPC), giving participants the technical foundation to scale their solutions efficiently, using the kind of infrastructure available through the Luxembourg AI Factory ecosystem.

Module 1: Introduction to AI and sustainability 

This module establishes the core vocabulary and frameworks required to evaluate AI initiatives from both a technical and sustainability perspective.
  1. Part 1 - Understanding AI through the AI lifecycle: Provides learners with a shared language and technical baseline covering problem definition, data collection, training, deployment, monitoring and improvement.
  2. Part 2 - Understanding sustainability: Broadens the perspective to ecological, social and economic lenses, and introduces the ESG and Sustainable Development Goals (SDG) decision frameworks professionals must use.
  3. Part 3 - Applying AI to sustainability challenges: Connects theory to concrete sectoral use cases such as electricity systems, mobility, agriculture and biodiversity.
  4. Part 4 - Deploying AI responsibly for sustainable value: Closes the progression by exploring implementation, governance, environmental footprint and responsible practice.

Module 2: Data-driven sustainability 

This module explores how sustainable AI succeeds or fails based on how well the underlying data represents the real system, the decision context and the conditions of deployment.
  1. Part 1 - Understanding sustainability data:Introduces major data types and explains how they are heterogeneous and decision-consequential.
  2. Part 2 - Building AI-ready spatial and temporal data: Details how Earth observation and spatiotemporal data are preprocessed, aligned and transformed into features and labels.
  3. Part 3 - Evaluating data quality for trustworthy sustainable AI: Examines how provenance, missingness, bias and uncertainty affect model reliability and fitness for purpose.
  4. Part 4 - Deploying and scaling Sustainable AI workflows responsibly: Introduces data spaces, data labs and infrastructure choices that balance performance with environmental responsibility.

Module 3: Foundations of machine learning 

This module focuses on the modelling layer, detailing how machine learning problems are framed and how datasets are prepared for reliable learning.
  1. Part 1 - Understanding machine learning problems: Introduces major machine learning paradigms and shows how problems are framed within an end-to-end workflow.
  2. Part 2 - Preparing data for reliable machine learning: Explains how data cleaning, scaling, feature construction and leakage-aware pipelines support trustworthy modelling.
  3. Part 3 - Predicting continuous and categorical outcomes: Presents foundational supervised-learning models to predict known targets using interpretable baselines.
  4. Part 4 - Discovering structure and predicting over time: Shows how to uncover patterns using unlabelled data and how to model temporal behaviour in evolving systems.
  5. Part 5 - Using neural networks and foundation models: Examines when neural approaches are useful and how larger pretrained models can be adapted.
  6. Part 6 - Evaluating models for trustworthy and efficient use: Concludes by assessing model quality, comparing alternatives fairly and balancing performance with efficient practices.

Module 4: HPC foundations for AI workloads 

This module focuses on the execution layer, showing how shared HPC systems are organised and how large-scale workflows can be run responsibly through Green AI practices.
  1. Part 1 - Understanding HPC for AI workloads: Introduces High-Performance Computing, distinguishing it from local or cloud environments and explains when sustainability workloads justify its use.
  2. Part 2 - Running AI workflows on HPC systems: Covers the practicalities of login nodes, compute nodes, schedulers, queues and job scripts for shared clusters.
  3. Part 3 - Building reproducible software environments: Focuses on the use of Python environments, containers, checkpointing and shared structures for collaborative practices.
  4. Part 4 - Managing data movement and performance: Addresses I/O bottlenecks by exploring storage layout, data locality, chunking and parallel loading pipelines.
  5. Part 5 - Using HPC sustainably and responsibly: Balances compute performance with cost, energy consumption and environmental impact through Green AI principles.

Module 5: AI for sustainable space and environment

Building on the geospatial data concepts from earlier modules, this module explores how AI processes complex inputs from satellite Earth observations and swarm coordination for Unmanned Aerial Vehicles (UAVs) to monitor ecological shifts. Learners will discover how advanced positioning systems and machine learning are deployed to track critical environmental metrics and manage natural resources dynamically.

Module 6: AI for optimisation, cybersecurity and smart Cities 

This module examines how machine learning and real-time algorithmic optimisation can drive resource efficiency across urban infrastructure while securing critical data systems. It bridges predictive modelling with resilient operations, demonstrating how AI can safely manage energy grids, mobility networks and sustainable smart city deployments.

Module 7: AI for sustainable finance 

Applying the complete AI lifecycle to the financial sector, this module illustrates how to process heterogeneous ESG (Environmental, Social and Governance) data to inform sustainable investment and policy decisions. Learners will use data-driven approaches to assess climate-related financial risks, ensuring that algorithmic models help align economic growth with long-term ecological stewardship.