AI engineer specializing in LLM systems and NLP, recently graduated and seeking full-time opportunities.
I recently completed my master’s degree in Computational Modeling and Simulation at TU Dresden, where I focused on NLP and LLM-based systems. My work centers on understanding model behavior, improving reliability, and making complex information more accessible.
Built a multi-threaded knowledge elicitation pipeline modeled after the GPTKB framework to construct a dense, domain-specific literary knowledge base. Utilizes graph structures to compute path-based recommendations and semantic similarity metrics over structural metadata rather than standard statistical correlation.
GitHub →Conducted a systematic study of false refusal behavior in LLM-based hate speech detoxification, revealing biases and providing guidance for improving fairness and robustness in safety-critical NLP systems.
Paper →Developed an LLM-based system for converting complex content into plain language. Evaluated outputs across BLEU, SARI, and readability metrics, and compared model behavior across GPT-4o and Llama.
Live → Paper → GitHub →Built a multi-label emotion classification model using RoBERTa with k-fold cross-validation and Optuna hyperparameter optimization. Analyzed model behavior with confusion matrices and SHAP.
Paper → GitHub →Designed and conducted a user study to generate gold-standard references for scientific text simplification, analyzing trade-offs in readability, accuracy, and consistency across prompting strategies.
Paper →Worked on LLM-based systems for scientific text simplification and NLP research. Developed and evaluated models across GPT-4, LLaMA, and BERT, and built structured datasets for large-scale experimentation.
Built and deployed production-grade data pipelines and infrastructure systems for enterprise analytics and ML workloads, with a focus on reliability, scalability, and automation using Python and Azure.
Developed embedded systems and simulation tools under strict performance constraints, and built testing and validation pipelines to improve system reliability and coverage.
M.S. in Computational Modeling and Simulation
Research focus: NLP, LLM systems, model behavior, and text simplification.
B.S. in Computer Science
ECIR 2025
Paper →