name: 'Efekan',
role: 'Backend & AI'
};
Hi, I'm
Efekan Çelik
Software Engineering student at Ankara University, focused on scalable backend systems, applied AI/ML, and distributed architectures. I build things that ship — from microservices to ML models to IEEE-published research.
01. About Me
I'm a software engineering student at Ankara University (GPA 3.25/4.00, expected 2027), with a passion for building scalable backend systems and applying AI/ML in real-world contexts.
My work spans Spring-based microservice architectures, PyTorch-powered deep learning models, and full-stack web applications. I've collaborated across teams at Turkcell, GreyCollar, and research labs — always shipping production-grade code.
Beyond engineering, I placed third nationally at TEKNOFEST 2025 in the 5G Positioning challenge, published in IEEE, and won first place at DATA LEAGUE'26 datathon out of 72 teams.
Education
02. Experience
Software Engineer Intern
Turkcell
- Built an Oracle DB-backed centralised configuration server for microservice environments, eliminating file-based config dependency.
- Designed an async push-based refresh mechanism propagating config changes to all bound microservices without restarts.
Software Engineer Intern
GreyCollar
- Designed and implemented test scenarios for an autonomous AI agent platform for workforce management.
- Measured LLM-based system output consistency and contributed to performance reporting pipelines.
- Gained experience with event-driven architectures and real-time data streaming principles.
Technical Lead
HSD Ankara University
- Managed community technical processes and led event infrastructure and presentation systems setup.
- Coordinated digital content management and event planning activities.
Software Engineer Intern
DivvyDrive
- Developed modular web components to enhance user interaction and interface responsiveness.
- Delivered a consistent, responsive UX across multiple screen sizes.
03. Projects
Spring Cloud Config — Oracle DB Demo
Enterprise-grade centralised config management for microservices. Async push-based config refresh propagates changes to all bound services without restarts. Built during Turkcell internship.
cvlangelo.com
AI-powered CV personalisation platform. Semantic search matches candidate experience to job listings. LangChain pipeline generates tailored CV text. Deployed and live.
Odyometre
Multidisciplinary clinical audiometry project across Ankara University faculties. Serial protocol between Java-based clinical software and Arduino/Proteus hardware layer.
KurbanLink
Full-stack livestock marketplace with listings, messaging, and vet appointment scheduling. Role-based auth, recommendation engine, AWS Lightsail deployment with CI/CD.
04. Publications & Competitions
Hybrid Deep Learning for 5G Positioning: Integrating Map-Derived Environmental Features with GAT-GRU-Attention
End-to-end 5G positioning architecture integrating map-based obstacle score features with a GAT-GRU-Attention (GGA) hybrid model. Achieved R² = 0.9805 and 21.26 m MAE in open-area scenarios.
IEEE XploreTEKNOFEST 2025 — 5G Positioning
Competed as "5Gordion" team member nationally. Developed a hybrid GAT-GRU-Attention model combining 5G signal measurements with map-derived environmental data.
T.C. Ministry of Industry and Technology · Sep 2025