YeonwooSung/MLOps

Provides hands-on implementations across the MLOps stack: serverless inference on AWS Lambda and SageMaker, distributed LLM fine-tuning with PEFT/QLoRA and FSDP, RAG pipelines, and model serving via Triton, Ray Serve, and FastAPI. Covers data versioning with DVC, Kubernetes orchestration, and LLM tooling including function calling and prompt engineering patterns. Includes practical examples for production deployments on AWS infrastructure and resource optimization guidance.