Adnan Kerem AKSOY

An enthusiastic and curious individual with a deep passion for learning and exploring new ideas. Demonstrates a creative approach to problem-solving and finds motivation in witnessing tangible results. Committed to continuous growth and thrives in challenging environments that foster the expansion of knowledge and skills.

• Certificate of Honor - Manisa Celal Bayar University
• IBM SkillsBuild, Agile Explorer - Powered by Agile at IBM, Collaborate Effectively, Enterprise Design Thinking Practitioner - IBM, Enterprise Design Thinking Co-Creator - IBM, Interpersonal Skills - IBM, Present With Purpose
• Elginkan Career, Electric-Electronic Measurements, Personality Analysis and Conflict Management, Healthy Life Management, Stress and Anger Management, Communication Skills

• Date of Birth: 08/01/97
• Driving License: B

• Collaborated with Prof. Dr. Sezai TASKIN on a university-level book introducing Mitsubishi Electric PLCs and Inverters. Created practical test sheets for students at Manisa Celal Bayar University's Electrical Equipment Laboratories, providing valuable documentation and troubleshooting guides for Electrical and Electronics Engineering students.
• Led the manufacturing traceability project as an Operational Technologies Engineer at Inci GS YUASA, overseeing design, implementation, and infrastructure development. Managed network cabling, EPC-OPC device configuration, and PLC module integration. Introduced networking solutions and a data logger for real-time production traceability.
• Developed various GAN (DCGAN, CGAN) and Neural Network (CNN, DCNN, AutoEncoders) applications for course projects during my ongoing ICT Master's degree at Padova.
• Conducted thesis research on the reconstruction of damaged old sound recordings using Generative Adversarial Networks (GANs), aiming to preserve sound originality during restoration. This study applied GANs to reconstruct and enhance corrupted audio signals, using both clean and corrupted samples for training. The process included preprocessing audio with MFCCs and STFT, testing different GAN architectures, and evaluating results through multiple comparison methods, including spectrogram analysis, listening tests, and five additional quantitative and qualitative metrics. This research provides a promising method for restoring damaged audio while maintaining its authenticity.