Personal Portfolio RAG Chatbot Case Study.

Overview

This workflow embeds CV data into a vector store and exposes a chat API that the portfolio frontend can call. Each question retrieves relevant chunks before response generation, and interactions are logged for analytics.

The Challenge

• Avoid hallucinated answers in a recruiter-facing assistant.
• Preserve context across follow-up questions in a natural chat flow.
• Keep conversation history queryable for review and improvement.
• Generate clear visibility into visitor intent over time.

My Solution

• Connected n8n webhook triggers with Pinecone semantic retrieval and OpenAI chat generation.
• Added window-buffer memory for session continuity.
• Persisted every Q&A pair in NocoDB for searchable history.
• Scheduled daily summary emails to monitor visitor question trends.