I build end-to-end, production-ready Generative AI systems. My focus is on the complete LLM lifecycle: from optimized fine-tuning (DPO, Unsloth) and advanced RAG architectures (GraphRAG, Self-RAG) to high-performance on-premise deployment (vLLM, FastAPI), robust evaluation (LangSmith), and security (LLM Guardrails).

My work bridges the gap between state-of-the-art research and scalable, production-grade applications.

• 🤖 Generative AI Engineering:

  • LLMOps & Production: High-performance inference with vLLM, API development with FastAPI, and containerization with Docker.
  • Model Fine-Tuning: Advanced model alignment using DPO (Direct Preference Optimization) and high-speed, memory-efficient training with Unsloth.
  • Evaluation & Observability: Tracing, monitoring, and evaluating RAG pipelines with LangSmith and custom evaluation frameworks.
  • Experiment Tracking: Full model lifecycle management and versioning with MLflow.
  • AI Security: Implementing LLM Guardrails to prevent prompt injection and ensure robust policy enforcement.

• ⛓️ Agentic AI & Advanced RAG:

  • Multi-Agent Systems: Architecting autonomous, collaborative AI agents using CrewAI and LangGraph.
  • State-of-the-Art RAG: Designing and implementing complex retrieval pipelines, including Agentic-RAG, Self-RAG, and Graph-RAG.
  • Core Frameworks: Deep practical experience with LangChain and LlamaIndex.

• 🎓 Academics & Research:

  • M.Sc. in Statistics & Computer Science: Currently researching Small Language Models and scalable RAG architectures.
  • B.Sc. in Statistics & Computer Science: Graduated 3rd in department (3.39/4.00 GPA).

I actively share my research, project breakdowns, and technical deep-dives on my Medium blog.

• 📄 Check out my articles: Medium Profile (@samet80)

📧 Email: abdulsamet80turkmenoglu@gmail.com