Hi there, I'm Yunus Emre Saidoğlu 👋

I am a 15-year-old open-source AI researcher and developer, proudly of Ahıska Turk heritage. I am the founder of @AhiskaAI, an independent research lab built to push the boundaries of Small Language Models (SLMs) and preserve cultural heritage through deep learning pipelines.

---

🚀 What I Do

• Pre-training & Fine-tuning: Training domain-specific SLMs from scratch (24M to 125M+ parameters).
• Data Engineering: Building, filtering, and curating clean synthetic datasets (Alpaca/ShareGPT formats).
• Tokenization: Experimenting with custom BPE tokenizers optimized for low-resource environments.
• Fail Forward: I actively document both my successful models and my training failures to help the open-source community learn.

---

🛠️ My Tech Stack

• Frameworks & Tools: Hugging Face Ecosystem, Transformers, Tokenizers, Datasets, PyTorch, Python
• Hardware: Powered by local compute & passion.