Skip to content
View sertacakalin's full-sized avatar

Highlights

  • Pro

Block or report sertacakalin

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
sertacakalin/README.md

Sertaç Akalın

Computer Engineer · AI Engineer Istanbul, Türkiye

I work on computer vision, applied AI, and backend systems. I enjoy technical problems that require building the evidence alongside the software: collecting data, designing evaluations, running experiments, measuring failure cases, and validating systems on real video.

Website · LinkedIn · Email

Research & Publications

  • Best Applied Research Paper — IDAP 2026 M. F. Şen, H. Gümüşkaya, and S. Akalın, “A deployable dual-mode vision framework for adaptive intersection control and traffic analytics.”
  • First-author paper accepted at IDAP 2026 S. Akalın and H. Gümüşkaya, “Computer vision-based detection of vehicles violating hatched road areas in Istanbul traffic.”

Both works are associated with the 10th International Artificial Intelligence and Data Processing Symposium, Istanbul, September 5–6, 2026. Extended journal manuscripts are in preparation for submission to Q1/Q2-indexed journals.

Selected Projects

An end-to-end computer vision system for detecting vehicles entering hatched road areas in fixed-camera Istanbul traffic videos.

  • Independently collected and annotated a dataset of 3,897 source frames and 9,353 exported images.
  • Fine-tuned YOLOv8m, reaching 0.773 mAP50.
  • Combined ByteTrack, polygon geometry, and multi-frame confirmation for spatial-temporal violation decisions.
  • Evaluated the full system on 11 real traffic videos, achieving 0.889 precision, recall, and F1.
  • Added severity scoring, evidence logging, and license-plate recognition in a Gradio interface; publicly released the dataset and code.

Palmystra

A shipped full-stack iOS application combining computer vision with a LangChain-based RAG pipeline to generate personalized readings from palm photos. Built with SwiftUI and a FastAPI backend, including authentication, rate limiting, and automated tests; deployed with Docker and Google Cloud Run.

Experience

  • Software Engineering Intern — FEV Türkiye · Jan–Jun 2026 Developed Python/FastAPI services and REST diagnostic endpoints for the SOVD project delivered to Jaguar Land Rover / Tata, working with engineering teams in Germany and India.
  • Java Backend Developer Intern — CodeFirst · Jan–Apr 2025 Developed Spring Boot microservice APIs with layered architecture, JWT authentication, and Spring Data JPA; used Git, Docker, Maven, Postman, and Swagger.

Technical Skills

Area Technologies & methods
Programming Python, Java
AI & Computer Vision PyTorch, YOLOv8, ByteTrack, OpenCV, model training and evaluation
Data & Research Dataset collection and annotation, Roboflow, quantitative evaluation, real-video field testing
Backend & Databases FastAPI, Spring Boot, REST APIs, PostgreSQL, MySQL, JWT
Tools & Deployment Git, Docker, GitHub Actions, Google Cloud Run, Postman, Swagger

Education

  • İstanbul Arel University — B.Sc. in Computer Engineering (English), 2022–2026
  • Associate Degree — Web Design and Coding, 2020–2022

Languages: Turkish (native), English (B2)

Resume

English CV · Türkçe CV

Popular repositories Loading

  1. hatched-area-violation-detection hatched-area-violation-detection Public

    Automated detection of vehicles illegally crossing hatched road markings using YOLOv8, ByteTrack, and trajectory-based severity scoring

    Python 1 1

  2. SpringPokeDB SpringPokeDB Public

    Java

  3. SpringSavePokemonDataBase SpringSavePokemonDataBase Public

    Java

  4. grayscale-ppm grayscale-ppm Public

    C++

  5. multi-threaded-image-pipeline multi-threaded-image-pipeline Public

    C++

  6. DormitorySystem DormitorySystem Public