Skip to content
View Berkaybbayramoglu's full-sized avatar
💭
👾
💭
👾

Highlights

  • Pro

Block or report Berkaybbayramoglu

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

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

Report abuse
Berkaybbayramoglu/README.md

header


Typing SVG



profile-views

🧠   About Me

"I build AI systems that are not just intelligent - but interpretable, reliable, and grounded in scientific rigor."

I'm an AI Researcher and Software Engineering undergraduate at Ankara University, currently a Research Assistant at the ANLAM-NET Research Lab and an Undergraduate Researcher at the CRADLE Lab, University of Michigan-Flint. My work spans both fundamental and applied deep learning - from the theoretical foundations (optimization dynamics, weight initialization, attention mechanics, uncertainty quantification) to real-world systems in computer vision, autonomous systems, geometric deep learning, trustworthy AI and reinforcement learning.

📄 For business inquiries, consulting, research collaborations, or recruitment purposes, you may request my CV for a detailed overview of my publications, research experience, and academic activities.

class Researcher:
    def __init__(self):
        self.name         = "Berkay Bayramoğlu"
        self.role         = ["Research Assistant @ ANLAM-NET Lab",
                             "Undergraduate Researcher @ CRADLE Lab, UM-Flint"]
        self.affiliation  = "Ankara University · Software Engineering (B.Sc., 2024–2028)"
        self.focus        = ["Computer Vision", "Autonomous Systems",
                             "Geometric Deep Learning", "Trustworthy AI",
                             "Reinforcement Learning", "Large Language Models"]
        self.principles   = ["Scientific Rigor", "Reproducibility",
                             "Interpretability", "Open Science", "Discipline"]

    def mission(self):
        return ("Advance the frontier of AI - building Computer Vision systems that "
                "perceive reliably under distribution shift, agents that turn perception "
                "into safe decisions, and Deep Learning "
                "models that are interpretable by design.")
  • 🔬 Currently exploring shift-robust visual perception, symmetry-aware graph neural networks, and reinforcement learning for autonomous systems
  • 📚 Active across fundamental and applied deep learning research
  • 🤝 Open to research collaborations, PhD discussions, and open-source contributions
  • 📄 Detailed CV available on request  -  for business inquiries, consulting, or recruitment

🛠   Technical Stack

🧪   Languages



🤖   AI / Deep Learning



📊   Scientific Computing



🧰   DevOps & Tools




⚛️   Quantum Computing



🏆   Competitions & Recognition

🥉Teknofest 2025 - 5G Positioning3rd Place · 343 teams · GNN+GRU+Attention hybrid · Turkcell datasets
🏅Teknofest 2025 - Turkish NLPTop 20 Finalist · 353 teams · Multi-step Reasoning · CoT · RAG · Function Calling
🥇YAZGİG'26 AI & Image Processing1st Place Winner · *Graphow* project
🎖UAB R&D Idea CompetitionSelected for official support under UAB Innovation Program

🔒 This is only a partial list — my major awards, research grants and publications are listed in my CV, available on request.


🤝   Let's Collaborate

I'm always open to research discussions, academic collaborations, and open-source contributions -
especially around Computer Vision, Geometric Deep Learning, Trustworthy AI, and Reinforcement Learning.




footer

⭐ "Science is a way of thinking much more than it is a body of knowledge." - Carl Sagan
Crafted with curiosity by @Berkaybbayramoglu

Popular repositories Loading

  1. NLP-Teknofest NLP-Teknofest Public

    LangNet-AU (Teknofest 2025 Turkish NLP, top 20 of 353): LangChain-based Turkish telecom agent with multi-step reasoning, RAG, function calling and optional speech I/O.

    Python 7 2

  2. Initialization-Method-Experiment-with-Neural-Net Initialization-Method-Experiment-with-Neural-Net Public

    Experimenting with different neural network weight initialization methods (e.g., Xavier, He, Orthogonal, and Poor Init) to visualize their effects on activation distributions, gradients, and traini…

    Jupyter Notebook 3

  3. numerical-gradient-optimization numerical-gradient-optimization Public

    This repository offers optimization algorithms implemented purely with NumPy. It features methods like SGD and NAG for function minimization, showcasing how basic linear algebra solves complex prob…

    Jupyter Notebook 2

  4. DeepFusion-Music-Genre-Classifier DeepFusion-Music-Genre-Classifier Public

    Multi-modal music genre classification in PyTorch, fusing CNN branches on mel-spectrograms with MLPs on MFCC, chroma and tempogram features.

    Python 2

  5. STEGANO STEGANO Public

    High-capacity image steganography engine in Rust with a Python CLI: adaptive quadtree decomposition, discrete ABC pixel selection and AES-256-GCM payload protection.

    Python 2

  6. Leaf-Disease-Classifier-Based-on-AI-Mobile-App Leaf-Disease-Classifier-Based-on-AI-Mobile-App Public

    Android app (Kotlin) that detects plant leaf diseases with an on-device CNN (ONNX) and gives treatment guidance via the Gemini API.

    Kotlin 1