ertugrul
@Ertugrul
Ertugrul Mutlu
Research-oriented computer engineering student working on ML systems and agentic...
About
Research-oriented computer engineering student working on ML systems and agentic AI. Interested in scalability, monitoring, and human supervision in real-world AI systems.
Language & Tools
Python, C++, SQL, Git/GitHub, LangGraph, PyTorch, PostgreSQL, ClickHouse, LLM APIs, Machine Learning, Agentic AI, AI Evaluation, Model Routing, and Observability.
Currently Exploring
Adaptive LLM routing, contextual bandits, neural representation dynamics, and reliable agentic AI systems. I'm particularly interested in how training history shapes neural networks and how we can build AI systems that are more efficient, interpretable, and reproducible.
Achievements
Working Student Researcher at Fraunhofer IAIS, focusing on applied AI and intelligent systems.
Author of research preprints exploring neural network behavior, training history, and representation dynamics.
Selected speaker at Python Summit 2026.
Creator of PromptLedger, an open-source prompt observability toolkit published on PyPI.
Developed MicroToolLM, an experimental language model focused on tool-use capabilities.
Author of research preprints exploring neural network behavior, training history, and representation dynamics.
Selected speaker at Python Summit 2026.
Creator of PromptLedger, an open-source prompt observability toolkit published on PyPI.
Developed MicroToolLM, an experimental language model focused on tool-use capabilities.
Fun Fact
I enjoy building AI systems from scratch, reverse-engineering software, and turning research ideas into working prototypes. When I'm not debugging neural networks, I'm probably playing guitar, volleyball, or exploring a new side project.
Random Dev Quote
"Two models can agree on every answer and still disagree on what they've learned."
Contributions
chevron_left