Internship topics

Get in touch
Internship topics

Federated Learning

MLOps for federated learning is an emerging field that seeks to address the challenges associated with deploying and managing federated learning systems in production. MLOps refers to the set of practices and tools used to streamline the development, deployment, and management of machine learning models. In the context of federated learning, MLOps involves managing the distribution of data and model updates across a decentralized network of devices or servers, ensuring the consistency and reliability of the system, and monitoring and optimizing performance. As a master thesis topic, MLOps for federated learning offers a unique opportunity to explore the intersection of two cutting-edge fields, and to develop innovative solutions to the challenges posed by privacy-preserving machine learning at scale.

AI act

The EU recently adopted the AI Act, which aims to regulate the development and use of artificial intelligence within the EU. The implementation of the AI Act will have far-reaching implications for businesses, organizations, and governments operating within the EU. To ensure compliance with the AI Act, it is crucial to educate stakeholders on the key provisions of the legislation and provide guidance on how to implement them. A workshop focused on the AI Act implementation would serve as a valuable forum for knowledge sharing and capacity building, bringing together experts and stakeholders to share insights and best practices for complying with the new regulations.


Sustainable AI is a critical topic for a master thesis, as the use of AI has the potential to significantly impact the environment and society. Sustainable AI refers to the development and deployment of AI systems that minimize their environmental impact, promote social responsibility, and ensure long-term sustainability. This involves considering the entire lifecycle of AI systems, from design to disposal, and taking steps to reduce their carbon footprint, promote ethical AI practices, and ensure equitable access. Research in sustainable AI can cover a wide range of topics, such as energy-efficient AI algorithms, responsible data management, and the ethical implications of AI systems. As AI becomes more pervasive in our lives, the importance of sustainable AI practices will only continue to grow, making it a critical and timely topic for a master thesis. (green AI, optimisation, edge) → can work on farad

Model fraud monitoring

Ensuring that a machine learning model is not tampered with or that no malicious behavior. With the increasing use of AI models in critical systems, it is important to ensure that these models are secure and free from malicious intent. To address this issue, several techniques can be employed, such as model encryption, anomaly detection, and adversarial robustness training. Model encryption involves encrypting the model parameters to prevent unauthorized access and manipulation. Anomaly detection techniques can be used to identify anomalous behavior and prevent malicious actors from exploiting vulnerabilities in the model. Adversarial robustness training involves training the model to resist attacks designed to manipulate or compromise the model's output. As a master thesis topic, research can focus on the development and evaluation of these techniques, as well as exploring new approaches to ensure the security and integrity of AI models.

In silico testing

Is a promising area of research that involves using computational models to simulate and test real-world systems. Digital twins are virtual replicas of physical systems that can be used to simulate their behavior and performance under different conditions. Machine learning can be applied to these digital twins to identify patterns and relationships in the data, allowing for more accurate predictions and improved performance. As a master thesis topic, this area offers an opportunity to explore the development and application of digital twin technology, the integration of machine learning algorithms with digital twins, and the evaluation of the effectiveness of these techniques in predicting the behavior and performance of real-world systems.


Solving constrained optimization problems with machine learning is an emerging field that involves leveraging machine learning techniques to solve optimization problems subject to various constraints. This approach is particularly relevant for problems where traditional optimization methods are computationally expensive or impractical. Machine learning models can be trained to learn the underlying structure of the problem and provide efficient solutions that meet the constraints. As a master thesis topic, solving constrained optimization problems with machine learning offers an opportunity to explore the intersection of two cutting-edge fields and develop new techniques that can help to enhance the efficiency and effectiveness of constrained optimization. Potential research topics in this area include developing new machine learning algorithms for constrained optimization, analyzing the accuracy and computational complexity of these algorithms, and applying them to real-world optimization problems in various domains.


Since the recent surge in popularity of Large Language Models (LLMs) such as LLaMa, LaMDA, Bloom, GPT, … and their chatbot-like implementations (e.g., ChatGPT, Stanford Alpaca, Google Bard, …) it becomes more and more important to implement these tools in a well-designed way. This means that we need to govern the behavior of these models to avoid and protect against undesirable behavior such as sensitive content (sex, drugs, violence, racism, harassment, …) and incorrect content (due to halucination). Master thesis topics include, among others, development of smart agents combining LLM capabilities with external information, developing guardrails to bound the behavior of the models and agents, building testing platforms to test the behavior and internal biases of these models and agents, automation of deployment of integrated applications, automation of feedback loops on such applications, …

Quantum computing

Quantum computing and kernel analysis is an emerging field that combines quantum computing with kernel methods, which are widely used in machine learning for feature extraction, classification, and regression. Quantum computing offers the potential for exponential speedup over classical computing, while kernel methods can effectively capture complex relationships and patterns in high-dimensional data. As a master thesis topic, research in this area can focus on developing and evaluating quantum kernel methods for various machine learning tasks, including image and speech recognition, natural language processing, and optimization. Potential research topics include exploring the potential advantages of quantum computing for kernel analysis, developing new quantum kernels, analyzing the computational complexity and accuracy of quantum kernel methods, and applying them to real-world problems. This area is at the forefront of research in quantum machine learning and has the potential to revolutionize the field of artificial intelligence.


We're here for you 🙌

Contact form

sign up to our weekly AI & data digest ❤️