Portfolio AI Assistant

Purpose-built to demonstrate practical AI agent architecture in a real production context. Rather than a generic chatbot, this assistant is scoped specifically to portfolio content: project documentation, architecture notes, and professional background. The RAG pipeline retrieves relevant content chunks from a vector store before passing context to the model, keeping responses grounded and auditable. The inference workflow is fully serverless: API Gateway routes requests through Lambda to the model API, with CloudWatch logging for every request and response. Safety controls constrain the prompt scope and prevent the assistant from generating content outside its defined boundaries. This is the same architecture pattern used in enterprise AI assistant deployments, applied at portfolio scale.