About
Experienced AI Platform Engineer with a strong background in building end‑to‑end machine learning systems, scalable backend infrastructures, and hybrid cloud deployments. Proven track record in deploying LLM and VLM models in production environments, designing real‑time data pipelines, and implementing CI/CD workflows. Adept at developing AI‑powered solutions for counterfeit detection, surveillance, and recom‑ mendation systems. Skilled in Kubernetes, Kafka, LangChain, and vector databases (Qdrant); hands‑on experience with AWS, GCP, and Azure. Combines deep AI knowledge with DevOps expertise to deliver robust, scalable, and observable AI platforms.
Work Experience
AI Platform Engineer at Counterfake AI ↗
Istanbul, Turkey
- Designed and developed AI systems for counterfeit detection on major e‑commerce platforms.
- Delivered protection solutions for both marketplaces (e.g., Trendyol) and global brands (e.g., Galatasaray, Beşiktaş).
- Deployed scalable AI models in Kubernetes with KEDA autoscaling and NVIDIA Triton Inference Server.
- Built high‑throughput data pipelines with Kafka and RabbitMQ; integrated LLM/VLM models (Florence, Qwen2.5‑VL) after benchmarking.
- Applied LoRA and PEFT for fine‑tuning; implemented RAG systems and agents via LangChain.
- Used Qdrant for vector search with image/text embeddings
- Designed observability stack with Grafana, ELK, and alerting mechanisms
- Led CI/CD pipeline automation and hybrid cloud architecture across AWS, GCP, Azure (Site‑to‑Site & P2P VPNs).
- Processed large datasets in Databricks using Parquet; utilized Hugging Face, PyTorch, TensorFlow.
Backend & AI/ML Developer | DevOps Engineer at Hybrone ↗
Istanbul, Turkey
- Participated in R&D projects, contributing to the completion of multiple TEYDEB projects with high evaluations.
- Authored and co-authored academic articles published in domestic and international journals.
- Developed chatbots using the Rasa framework, reducing technical service workload by 30% and providing 24/7 user support.
- Implemented AI-driven features such as voice and camera support for mobile apps.
- Conducted fraud detection in retail using CCTV cameras and YOLO models, achieving over 90% accuracy.
- Worked on Bluetooth and UWB-based indoor positioning systems.
- Developed facial recognition systems with custom facenet-based hybrid models, integrating them into the Hybrone platform.
- Built hybrid recommendation models for personalized user experiences.
Backend Developer at Biges ↗
Istanbul, Turkey
- Developed the backend for the company's B2B platform using Django Rest Framework, integrated with Canias ERP.
- Managed ERP interactions to provide real-time information on dealer statuses, product stocks, and technical services to end users.
- Utilized PostgreSQL for database management, Celery for task queuing, Redis for caching, and Elasticsearch for search indexing.
- Employed S3 and GCP for storage, and Docker for containerization.
- Ensured the platform's active usage by over 3000 dealers, introducing an advanced, user-friendly B2B system in the sector.
Education
Master Degree at Fırat University
Elazig, Turkey
Bachelor Degree at Fırat University
Elazig, Turkey
Writing
Market Fraud Detection with Image Processing ↗
International Conference On Emerging Sources In Science
Face Identification in Security Systems ↗
Cukurova 9th International Scientific Researches Conferance
Product Recommendation System Based on AI ↗
RDCONF International Conference on Design, Research and Development
Certifications
Certificate of Participation
IKSAD Institute
RDCONF2021 Certificate of Participation ↗
Orclever Science&Research Group