Simon Ostermann

Simon Ostermann

NLP & AI Researcher @ DFKI. Mechanistic interpretability and multilingual NLP, with a focus on low-resource languages.

Students

I'm happy to collaborate with and advise the following PhD students:

PhD Students

Tatiana Anikina

Topic: Efficient Adaptation of Language Models for Low-Resource NLU

Cennet Oguz

Topic: Multimodal Processing of Procedural Knowledge

Tanja Baeumel

Topic: Mechanistic Interpretability of LLMs

Yusser al Ghussin

Topic: Mechanistic Steering of LLMs in Multilingual and Multicultural Aspects

Daniil Gurgurov

Topic: Interpretable NLP for Low Resource Languages

Ivan Vykopal KInIT

Topic: Multilingual Low-resource NLP

My team and I are happy to supervise the following MSc theses:

Ongoing MSc Theses

Investigating SAEs to Improve Clustering and Classification

Aarushi SinghalMSc

Improving Language Representation in LLMs with Language Tokens

Kaviya RavichandranMSc

Mechanistic Understanding of LLM Adaptation

Khondoker Ittehadul IslamMSc

If you are interested in writing your MSc thesis with my E&E group, please send me a mail with a short topic outline, a CV and a transcript of records. Please note that we get many such requests and cannot collaborate with everybody that is interested.

Finished MSc Theses (13)

Localizing Modular Translation Strategies in Multilingual Large Language Models

Mikhail Sonkin·2026Advisor

Efficiently Injecting Knowledge into Large Language Models: An Investigation of Approaches for Extremely Large Label Spaces in the Library Domain

Katja Konermann·2025Advisor

Reading Between the Layers: Investigating Impacts of Adapters and Few-Shot-Learning on the Internal Latent Language of LLMs

Gregory Charles Shook·2025Advisor

Bridging the Language Gap: Developing Urdu and Roman-Urdu Information Retrieval Systems

Umer Tariq Butt·2025Reviewer

Mitigating Hallucinations on a Multi-Source Summarization Setting

Lucas Lage·2025Advisor

External Knowledge-Driven Adaptation of Multilingual Language Models for Low-Resource Language Enhancement

Daniil Gurgurov·2025Advisor

On the Efficacy of Language Adapters for Cross-Lingual Transfer in English-Centric LLMs

Julian Schlenker·2025Advisor

Analysis of Calibrated Confident Text Classification

Bangyao Tang·2024Reviewer

Self-Supervised Multimodal Representation Learning for Diagram Understanding

Akshai Joshi·2024Reviewer

Fast and Efficient Structured Pruning of LLMs with Gradient-Based Meta-Mask

Konstantin Chernychew·2024Advisor

Extreme model compression for large-scale transformer-based language models

Gokul Srinivasagan·2023Advisor

Cross-lingual and Cross-domain Knowledge Transfer for Sequence Labeling Tasks

Daria Fedorova·2023Advisor

Investigating Instantiations of Script Structures in Narrative Texts

Hannah Seitz·2016Advisor