Veli Karakaya
Undergraduate Researcher, Bilkent University
Software Analytics · AI4SE · LLM Evaluation
I am an undergraduate Computer Engineering student at Bilkent University, Türkiye, researching AI for Software Engineering (AI4SE), Software Analytics, and the evaluation of large language models. I am particularly interested in how AI systems can be effectively integrated into software development processes, including code review, program analysis, and developer support tools, with an emphasis on evaluation, reliability, and real-world applicability.
I work at BILSEN Lab, Bilkent University, under Assoc. Prof. Eray Tüzün, where I research software analytics and LLM-based evaluation of automated code review. I am also a visiting researcher at SeSa Lab, University of Salerno, under Assoc. Prof. Fabio Palomba, where I study code ownership dynamics in the context of LLM-based coding agents.
I have contributions accepted at EASE 2026 and ASE 2026 (ATIQSER), and I plan to pursue advanced research at the intersection of LLM evaluation and software engineering.
Publications
Experience
- Researching code ownership dynamics in the context of LLM-based coding agents.
- Researching software analytics, ML for software engineering, and LLM evaluation in automated code review contexts.
- Designing evaluation methodologies to identify limits of automated metrics in code review quality assessment.
- Co-authored three papers accepted at EASE 2026 (Industry and Vision Tracks) and ASE 2026 (ATIQSER).
- Designed and developed a full-stack mobile application using Flutter and PostgreSQL with authentication and ownership-based access control.
- Integrated 3D geographic modules for interactive visualization of cultural and strategic sites.
- Implemented AI/ML-based recommendation engine to suggest nearby points of interest based on user context.
- Built deduplication logic and real-time GPS centering to ensure data integrity and UX quality.
Projects
- Automated internal paper review workflow: reviewer assignment, deadline tracking, email notifications, and completion confirmation.
- Implemented accept/decline flow with reason logging, deadline extension requests, and workload visibility dashboard.
- Built a location-based memory sharing app on Google Maps with media content and 3D location modelling.
- Integrated AI-based recommendation logic to suggest nearby points of interest based on user context.
- Developed a desktop system for blood donation matching based on blood type compatibility and scarcity metrics.
- Implemented real-time donor notifications; designed and fully normalized the database schema.
- Designed and implemented a fully normalized relational database for restaurant operations (reservations, orders, staff, menu).
- Wrote complex reporting queries; applied indexing and query optimization on high-frequency access patterns.
Presentations
- Co-presented (with U. B. Torun), "Understanding the Limits of Automated Evaluation for Code Review Bots in Practice." EASE 2026, Industry Track, Glasgow, UK, June 2026.
- Co-presented (with U. B. Torun), "Evaluation of LLM-Based Software Engineering Tools: Practices, Challenges, and Future Directions." EASE 2026, Vision Track, Glasgow, UK, June 2026.
- Attendee, 18th International Summer School on Software Engineering (ISSSE 2026), University of Salerno, Italy, June 2026.
Education
- CS390 Individual Research Study (A): End-to-end research on LLM-based evaluation of automated code review (literature review → methodology → experiments → reporting).
- CS319 Object-Oriented Software Engineering (A+): Developed a full-stack research lab management tool supporting hallucination detection in academic papers, AI-assisted paper review, reviewer assignment, deadline tracking, and venue recommendations.