I design, build, and optimize production-oriented AI systems focused on speech processing, local large language models, and reliable API integration. I develop Turkish or multilingual transcription pipelines using Whisper, WhisperX, Faster-Whisper, and Silero VAD; deploy containerized LLM services with Docker and vLLM; create structured-output and retrieval workflows with FastAPI, Qdrant, and SQL; and integrate computer-vision components using YOLO, OpenCV, and OpenVINO. My approach emphasizes measurable performance, resource planning, observability, and dependable outputs. Available deliverables include architecture, prototyping, deployment, debugging, optimization, documentation, and end-to-end integration on Linux/CUDA environments. Each engagement is scoped around real operational requirements.