Developed a tool to evaluate LLMs such as GPT, BERT, LLaMA on various tasks with few-shot learning support.
Implemented custom evaluation metrics for accuracy and perplexity.
Occupation Prediction from Twitter Data
Developed an NLP project to analyze and classify social media data.
Used ML models such as Logistic Regression, SVM, Random Forest, CNN, and RNN to explore the impact of dataset size on model performance, achieving 89% accuracy in classification tasks.
Publications
Refereed Journal Publications
T. İzdaş, H. İskifoğlu, and B. Diri, “Occupation Prediction from Twitter Data", DEUFMD, vol. 27, no. 80, pp. 267–271, 2025, doi: 10.21205/deufmd.2025278013.
Certifications
Red Hat System Administration I & II -
RH124 &
RH134
Red Hat OpenShift I: Containers & Kubernetes -
DO180