AI Engineer building autonomous multi-agent systems and the production platforms that run them, with hands-on experience across the Model Context Protocol (MCP), RAG, and LLM orchestration. Outside of work, I'm independently building a real-time multiplayer mobile game. My work sits at the intersection of AI, backend architecture, and full-stack product engineering.
Here is a showcase of my work, from autonomous AI agents and real-time data pipelines to full-stack applications.
| Project | Core Technologies |
|---|---|
|
Next-Gen Support Agent Blueprint
An end-to-end, self-correcting customer support agent that uses a reasoning loop (LangGraph), RAG, and intelligent tools to handle complex user queries. |
|
|
AIDA - Autonomous Incident Diagnostic Agent
An end-to-end MLOps system for an autonomous SRE agent that diagnoses production alerts, learns from feedback, and improves over time via a fine-tuning pipeline. |
|
|
Nebula - Real-Time Recommendation Engine
A real-time recommendation engine that reacts to user interactions in seconds, using a stateful stream processor (Bytewax) and a large-scale vector database (Milvus). |
|
|
CodeScribe - AI Code Review Agent
An autonomous AI agent that integrates with GitHub to provide intelligent, RAG-powered code reviews and interactive refactoring suggestions on Pull Requests. |
|
|
Real-Time Fraud Detection System
An end-to-end streaming data pipeline that uses Apache Flink and a stateful ML model to detect fraudulent transactions in real-time. |
|
|
GitOps CI/CD Pipeline for Kubernetes
A complete GitOps workflow using Argo CD to manage the deployment of a microservices application onto Kubernetes with a multi-environment promotion strategy. |
|
|
AI-Powered Automated Attendance System (AttendAI)
A high-performance, distributed AI system for real-time facial recognition, featuring a full-stack observability dashboard. |
|
|
Sakarya University AI Chatbot (RAG System)
A research-oriented RAG chatbot built to provide accurate answers from a large, domain-specific knowledge base. |
|
|
LanDrop - Local Network File Transfer
A cross-platform desktop application for seamless file transfer over a LAN using Python sockets and Zeroconf for automatic device discovery. |
|
- AI & Agents: LangChain / LangGraph, Model Context Protocol (MCP), Microsoft Agent Framework, Microsoft Foundry, RAG, Vector Databases (Qdrant, FAISS), Local LLMs (Ollama, llama.cpp, vLLM)
- Backend: Python (FastAPI), .NET (ABP Framework), Node.js (Fastify), PostgreSQL, Redis, ClickHouse
- Frontend & Mobile: TypeScript, React.js, Next.js, React Native (Expo)
- DevOps & Cloud: Docker, CI/CD (GitHub Actions), Prometheus / Grafana, Azure (Azure OpenAI), Cloudflare