Professional Summary
I’m an AI Engineer based in Istanbul. As a recent graduate, I am highly focused on building scalable, testable, and maintainable systems. During my university years, I actively participated in hackathons and hands-on engineering challenges, which taught me how to deliver solutions under tight deadlines. This practical approach has allowed me to build and successfully deploy live projects. My technical foundation spans full-stack development and AI, with a special interest in computer vision and UAV systems. I am a team-oriented engineer, always eager to learn from experienced peers while contributing to real-world, production-ready software
Technical Skills
Experience
- Developed a sensor fusion module for GNSS-denied environments, supporting autonomous flight and safe Return-to-Home behavior when satellite navigation is unavailable.
- Built the module to meet defence-grade reliability and real-time performance requirements within a safety-critical autonomous system.
- Developed a real-time mobile control interface for the Ground Control Station (GCS), giving operators live visibility into vehicle status and telemetry.
- Contributed to data preparation, model training, and inference pipelines for UAV-based object detection and tracking systems.
- Worked on preprocessing and structuring datasets used to train and evaluate detection models under real flight conditions.
- Developed a real-time mobile interface for the Ground Control Station to monitor the vehicle's live status and telemetry, work later extended into the Software Engineer role.
- Completed an 8-month industry research project under the SAYZEK program, supervised by Assoc. Prof. Dr. Caner Özcan (Karabük University) with industry mentorship from Dr. Cevahir Çığla (ASELSAN).
- Evaluated CNN and R-CNN based architectures against the project's technical requirements and selected the approach that best fit the target use case.
- Built an end-to-end training pipeline covering data collection, preprocessing, and augmentation for the selected model.
- Built a real-time visual perception system using OpenCV and PyTorch to support the drone's autonomous mission logic.
- Deployed the resulting models as edge AI on embedded Linux hardware, working within the compute and power constraints of an onboard system.
- Established TCP/UDP communication between the drone and the Ground Control Station and coordinated the team responsible for the software stack.
- Developed a web-based medical image annotation platform with .NET and Vue.js, under the supervision of Prof. Dr. Hakan Kutucu, used by radiologists to label 28,800 images.
- Designed the MySQL data model behind the platform, structuring it for consistent, high-performance access as the annotation dataset grew.
- Developed backend features in .NET for inventory and warehouse modules, including SQL queries and RESTful APIs.
- Worked within an existing Git-based team workflow, participating in code reviews and adapting to established development practices.
Projects
Flight control and computer vision simulations built for tactical, autonomous UAVs.
Developed an end-to-end autonomous navigation and computer vision suite for fixed-wing UAVs. Designed systems for real-time moving target tracking, precision payload delivery via geometric shape recognition, and automated QR-code-based dive maneuvers. Successfully integrated flight control software with computer vision pipelines under simulated real-world physics constraints.
Backend for an application that detects disease from real radiographic images.
Built a robust backend architecture for a disease diagnosis web application. Engineered secure and efficient data pipelines to process and annotate real radiographic images, focusing on system reliability and clean API design for seamless frontend integration.
End-to-end machine learning pipeline that pulls live data from the web.
Engineered a complete machine learning pipeline to predict housing rents in Istanbul. Extracted raw data via web scraping, performed advanced feature engineering, and evaluated multiple regression algorithms. Achieved the highest prediction accuracy using XGBoost and Random Forest models by handling outliers and optimizing feature sets.
Full-scope, role-based software built for production and warehouse management.
Designed and developed a full-stack Enterprise Resource Planning (ERP) application for manufacturing management. Implemented modules for real-time stock tracking, warehouse operations, and role-based user authorization (Admin, Production Planning, IT), ensuring a scalable and maintainable codebase.
Testing infrastructure demonstrating the ability to write high-quality, testable code.
Created a comprehensive testing repository to validate software reliability across different tech stacks. Implemented black-box testing methodologies, in-memory database integration tests, and unit testing architectures to ensure high code coverage and fault-tolerant system design.