Contact Information
- Email [email protected]
- Phone +90 531 340 7939
- LinkedIn linkedin.com/in/dogukannparlak
- GitHub github.com/dogukannparlak
- Portfolio dogukanparlak.com
Professional Summary
4th-year Computer Engineering student with two internships in backend development and applied AI. At Hyperever, trained real-time object detection models with YOLOv11 and OpenCV; at YoSoft, built Flask/MySQL-based REST APIs. In personal projects, compared symbolic (FOL) and LLM-based multi-agent systems and reached 75% accuracy with machine learning models.
🚀 Technical Skills
Languages
Python
JavaScript
SQL
Machine Learning / AI
PyTorch
scikit-learn
XGBoost
OpenCV
YOLOv11
Backend & API
FastAPI
Flask
SQLAlchemy
REST API
Tools
Experience
Hyperever (ITU ARI Technopark)
Computer Vision Intern
- Built and annotated a dataset of 42,000+ images using Label Studio and COCO-based annotation pipelines to support large-scale single-class object detection training.
- Trained and optimized real-time object detection models using YOLOv11, OpenCV, and Ultralytics, improving target-class discrimination against visually similar objects.
- Developed data preprocessing and fine-tuning pipelines to enhance model robustness and deployment readiness for integration into a robotic dog platform.
Pasific Bilişim / YoSoft
Software Development Intern
- Designed and developed a WhatsApp-based bulk messaging system to streamline operational communication between a technology company and its pharmacy client network.
- Built a Flask-based REST API with MySQL integration and session-based authentication, supporting multi-profile usage and dynamic message handling.
Education
🎓 Computer Engineering (English)
Aydın Adnan Menderes University · 4th Year
September 2022 - September 2027
Languages
- Turkish Native
- English Language of Instruction (English-Medium Program)
Open Source Contributions
instagram_monitor
Merged contribution adding a fully offline follow-relationship analysis feature (mutual/non-mutual/fan detection from locally cached follower data), accessible via both a CLI flag and a web dashboard modal, verified with Pytest tests (PR #121, merged).
DynamicNotch
Merged contribution adding full Turkish localization support (UI, settings, Info.plist permission strings), verified by dedicated tests (PR #189, merged).
ArcLauncher
Merged contribution adding full Turkish localization support for an Android TV launcher app (PR #34, merged).
Selected Projects
🎮 Digital Game Marketplace
A full-stack digital game storefront inspired by Steam and Epic Games, with purchase flow, a developer portal, and an admin panel backed by a role-based REST API.
🆘 Disaster Rescue Multi-Agent System
A multi-agent disaster rescue coordination system built for cooperative response simulation, with two repository components: a symbolic FOL edition and an LLM-based Ollama edition. The FOL edition achieved 100% deterministic results across all runs (task matching, completion, cost ratio 1.0), while the LLM edition produced results fluctuating between 80-100% on the same seed, concretely demonstrating LLM non-determinism.
🎬 Movie Rating Predictor
A machine learning project for predicting movie ratings using data preprocessing, feature engineering, and model evaluation. On 1,312 movies, Gradient Boosting classification reached 75% accuracy (ROC-AUC 0.81), while SVR regression achieved a MAE of 0.93 (on a 1-10 scale).
📬 Mail Sender Extractor
A Gmail inbox-cleaning web app: scans your inbox, groups emails by sender with counts, and lets you bulk-trash or one-click unsubscribe from selected senders - plus a CLI-only mode for CSV export.
Want to Learn More?
Get in touch to meet me, talk about my projects, or collaborate.