Muruvvet Bozkurt

Automated IT processes with text analysis, clustering, and NLP, cutting processing time by 83.3% and saving €30,000 annually. Developed a LangChain-based RAG chatbot deployed on OpenShift, using FAISS and internal knowledge sources to automate IT support.

Built a custom neural network and optimized various machine learning models for accurate meal predictions based on patient order history, enhancing assignment and satisfaction. Applied ARIMA/SARIMA and sequential models for future patient presence prediction, integrated with BigQuery.

Created an explainable object detection model for autonomous driving safety using autoencoders and Bayesian causal models in PyTorch. Investigated advanced models like Faster R-CNN, YOLO, and Mask R-CNN.

Proposed a model to detect facial expressions in children with limited datasets, creating domain-invariant models by reducing shifts between child and adult facial features. This approach was inspired by GAN, WGAN, and gradient reversal techniques.

Developed a React-based interface for the MOST Project to upload Excel files, read data, and send it to designated services. This project streamlined data handling processes.

Developed DL models to classify NIR spectrometer images of rocks and minerals, and created autonomous traversal algorithms. Designed the Rover’s state machine using Python and ROS, implementing a custom motion planning library with lidar data for desert navigation.

Simulated the motion of knees and elbows using textile-based sensors with Unity. This involved creating virtual environments to test sensor data.

Pursuing a Master of Science degree in Informatics. Focusing on advanced topics in computer science and information technology.

Grade: 1.7 in German

Completed a Bachelor of Science degree in Computer Engineering. Final project focused on Facial Expressions Detection of Children with Hearing Impairment, which was published as a paper.