I study computer engineering at Manisa Celal Bayar University. I'm focused on AI/ML and deep learning, and I like building things that run locally, from small language models to RAG agents on edge hardware.
- Building maidere, my private, local-first AI agent.
- Built study-assistant for the Nebius × NVIDIA Global AI Hackathon.
- Built whisperboard at the MLH × Midnight Hackathon.
| Project | What it is | Built with |
|---|---|---|
| maidere | Privacy-first AI agent for your own machine, with long-term memory, subagents, vision and sandboxed tools, running on local models with optional cloud scaling | Python, FastAPI, LangGraph, Ollama, SQLite |
| study-assistant | Academic tutor that answers from your notes or the web, with citations and SymPy-checked math | Python, TypeScript, NVIDIA Nemotron, ChromaDB |
| whisperboard | Anonymous feedback board with zero-knowledge proofs on the Midnight Network | TypeScript, React, Compact |
| SovereignNode-AI | Air-gapped edge node that turns factory sensor data into maintenance diagnostics with a local SLM | Java, Spring Boot, FastAPI, React |
| SolQuant | Local-first RAG agent stack tuned for low-VRAM GPUs | Python, Java, llama.cpp, MongoDB |
| small-language-model | Transformer framework for training 1B–8B models with GQA, RoPE and quantization | Python, TensorFlow |
AI / ML Python, TensorFlow, NumPy, LangGraph, Ollama, llama.cpp, RAG, ChromaDB
Backend FastAPI, Java, Spring Boot, Node.js, MongoDB, SQLite
Frontend React, TypeScript, Tailwind CSS
Also C, Dart