Ramazan Bahrami

Albrecht-Thaer-Weg 24f

37075 Göttingen, Germany

I am a Doctoral Researcher (Expected to grduate by the end of 2026) at the University of Göttingen specializing in Natural Language Processing (NLP), Graph Machine Learning, and Generative AI . My research centers on enhancing the reasoning capabilities of Large Language Models (LLMs), relation extraction using advanced architectures, and graph-based data processing.

With over 5 years of software engineering experience alongside my academic research, I bridge the gap between theoretical AI foundations and practical software development. I have hands-on expertise building AI-driven solutions—ranging from graph query processing to LLM agents and educational technology platforms.

Research & Highlights

Reasoning in LLMs: Developed novel approaches like GraphMapping to evaluate and benchmark reasoning capabilities across large language models.
Relation Extraction: Designed capsule-network architectures (Re-Representation) that achieve state-of-the-art performance in sentential relation extraction.

news

Sep 07, 2026	Our paper, Evaluating LLMs Thinking Speed, titled : “Graph Mapping and LLM Thinking Speed”, was accepted at the 17th International Conference on Natural Language Generation (INLG) in the Netherlands. Read it here.
Jun 15, 2025	Our paper “Re-Representation in Sentential Relation Extraction with Sequence Routing Algorithm” has been accepted at ICNLSP 2025! Check out the code on GitHub or read our publications.

selected publications

ICNLSP

Re-Representation in Sentential Relation Extraction with Sequence Routing Algorithm

Ramazan Bahrami and Ramin Yahyapour

In Proceedings of the 8th International Conference on Natural Language and Speech Processing (ICNLSP-2025), 2025

Bib Code

@inproceedings{bahrami2025rerepresentation,
  title = {Re-Representation in Sentential Relation Extraction with Sequence Routing Algorithm},
  author = {Bahrami, Ramazan and Yahyapour, Ramin},
  booktitle = {Proceedings of the 8th International Conference on Natural Language and Speech Processing (ICNLSP-2025)},
  publisher = {Association for Computational Linguistics},
  pages = {315--327},
  year = {2025},
}

ICNLP

Multihop Question Answering, A Topological Approach for Graph Generation

Ramazan Ali Bahrami and Ramin Yahyapour

In 2024 6th International Conference on Natural Language Processing (ICNLP), 2024

Bib

@inproceedings{bahrami2024multihop,
  title = {Multihop Question Answering, A Topological Approach for Graph Generation},
  author = {Bahrami, Ramazan Ali and Yahyapour, Ramin},
  booktitle = {2024 6th International Conference on Natural Language Processing (ICNLP)},
  pages = {770--774},
  year = {2024},
}

Efficient processing of SPARQL queries over GraphFrames

Ramazan Ali Bahrami, J. Gulati, and M. Abulaish

In Proceedings of the International Conference on Web Intelligence (WI ’17), 2017

Bib

@inproceedings{bahrami2017efficient,
  title = {Efficient processing of SPARQL queries over GraphFrames},
  author = {Bahrami, Ramazan Ali and Gulati, J. and Abulaish, M.},
  booktitle = {Proceedings of the International Conference on Web Intelligence (WI '17)},
  pages = {678--685},
  publisher = {ACM},
  year = {2017},
}