RAG Pipeline & Podcast Summary Studio
Nico Belazaras
Developed a Retrieval-Augmented Generation (RAG) pipeline and a companion Podcast Summarization Studio, showcasing how businesses can turn unstructured content into searchable knowledge and concise insights.
Highlights:
Document ingestion & search: PDFs and text files processed into embeddings with pgvector, enabling semantic Q&A with citations.
AI chatbot UI: Built with Next.js, streaming responses from OpenAI/Anthropic models, complete with source highlighting.
Podcast summarization: Upload audio episodes, transcribe, and auto-generate structured summaries for quick review.
Extensible design: Authentication-ready and deployable on Vercel + Supabase for quick iteration.
Impact:
Demonstrated how AI can reduce research time, streamline knowledge access, and transform long-form media into actionable insights.
Tech Stack: Next.js, Supabase/pgvector, OpenAI/Anthropic APIs, TypeScript, Tailwind CSS.
Documents to Q&A Chatbot
Podcast to Blog Posts
Like this project
Posted Sep 30, 2025
Retrieval-Augmented Generation (RAG) pipeline & Podcast Summarization, showcasing how to turn unstructured content into searchable knowledge and insights.
