Skip to content
YusufKorkmazyiğit
TR

Yusuf Korkmazyiğit

Computer Engineering Student · AI and Optimization

yusuf.korkmazyigit@gmail.comyusufkorkmazyigit.comgithub.com/yusufkorkmazyigitlinkedin.com/in/yusufkorkmazyigit

Experience

  1. Oct 2023 — present

    Undergraduate Researcher (TÜBİTAK 2247-C fellow)

    Erciyes Nötrino Araştırma Grubu · Kayseri

    • Selected for the TÜBİTAK 2247-C fellowship to work on "Improving Neutrino Physics Analyses with AI Techniques".
    • Built deep learning models in PyTorch for high-energy physics data, improving signal detection efficiency.
    • Built data preprocessing pipelines for large-scale neutrino simulations together with senior physicists.

    Python · PyTorch

  2. Jul 2026 — Aug 2026

    Software Intern — IT / Automation

    POSCO Assan TST Çelik San. A.Ş. · İzmit, Kocaeli

    • Ran an end-to-end data analysis project on coil sensor data to identify process parameters driving surface defects on the APF (annealing, pickling and finishing) line — SQL extraction, cleaning, label engineering and EDA.
    • Compared logistic regression, decision tree, Random Forest and XGBoost; caught row-level data leakage and switched to coil-level GroupKFold validation (test ROC-AUC 0.87), explaining feature importance with SHAP.
    • Showed with before/after analysis that the defective-record rate in live production dropped from 33.0% to 10.0% after a parameter change.
    • Trained a YOLOv8n-seg segmentation model on a 12,568-image steel surface dataset and presented results in a Streamlit app.

    Python · SQL · XGBoost · SHAP · YOLOv8 · Streamlit

  3. Feb 2024 — Jul 2026

    Principal Investigator (undergraduate research grant)

    TÜBİTAK 2209-A · Kayseri

    • Led the TÜBİTAK-funded project "Optimizing Machine Learning Models for Cherenkov/Scintillation Separation with the Artificial Bee Colony (ABC) Algorithm".
    • Tuned model hyperparameters with ABC, clearly improving photon classification accuracy.
    • Ran a comparative analysis of metaheuristic algorithms and reported the findings in a technical project report.

    Python · Scikit-learn · NumPy

Education

  1. Sept 2022 — present

    B.Sc. Computer Engineering — GPA 3.19 / 4.00

    Erciyes Üniversitesi · Kayseri

    • Expected graduation: 2027. Capstone: KOVAN — interpretable algorithm discovery with LLM-guided ABC.
    • High honors for three semesters; term GPAs 3.72 (Spring 2024–25) and 3.70 (Fall 2025–26).
    • Relevant courses: deep learning, data mining, scientific programming, probability and statistics, linear algebra, numerical methods, data structures and algorithms.
  2. Feb 2025 — Jul 2025

    Erasmus+ Exchange Student — term GPA 3.72 / 4.00

    Częstochowa Teknoloji Üniversitesi · Częstochowa, Polonya

    • 30 ECTS in English: image processing and recognition, computer architecture, databases and data warehousing, Java, web scripting, numerical methods.
    • Course projects on image enhancement and face recognition with OpenCV and TensorFlow.

    OpenCV · TensorFlow · Java

  3. Sept 2021 — Jun 2022

    B.Sc. Computer Engineering — GPA 3.86 / 4.00

    Selçuk Üniversitesi · Konya

    • Ranked first in the department's first-year class; transferred to Erciyes University in 2022.

Projects

  • KOVAN — Interpretable algorithm discovery with LLM-guided Artificial Bee Colony — my capstone project.
    Python · LLM · Artificial Bee Colony · Optimization
  • abcopt — Canonical Artificial Bee Colony optimization for Python with a scikit-learn compatible hyperparameter search. Published on PyPI.
    Python · NumPy · scikit-learn · PyPI
  • ABC-FedAvg — Optimizing federated learning hyperparameters on distributed medical data with Artificial Bee Colony (MedMNIST).
    Python · PyTorch · Federated Learning · Artificial Bee Colony
  • Steel surface defect detection — Defect segmentation with YOLOv8n-seg on a 12,568-image steel surface dataset, served through a Streamlit app.
    Python · YOLOv8 · Ultralytics · Streamlit

Skills

Programming
Python (ileri), Java, SQL, JavaScript
AI / ML
PyTorch, Scikit-learn, XGBoost, SHAP, YOLOv8, OpenCV, NumPy, Pandas
Methods
Metaheuristic optimization (ABC), Hyperparameter optimization, Federated learning, Image segmentation, Explainable AI
Data
PostgreSQL, MySQL, Oracle, Data warehousing
Tools
Git / GitHub, Linux, Jupyter, Streamlit, PyPI packaging
Languages
Turkish (native), English (C1)

Community

  1. Sept 2022 — present

    Actor and Backstage Crew

    Erciyes Üniversitesi Tıp Fakültesi Tiyatro Topluluğu · Kayseri

    • Played Mateh Usta in Aziz Nesin's "Biraz Gelir misiniz?" and worked backstage on productions.
    • Staged plays for university events with a cast of more than 20.