Build a Multi-Agent AI System for Long-Form Economics Book WritingSearching

AI Developer is needed in Mumbai, India.

Build a Python-based, stateful multi-agent workflow for researching, outlining, drafting, and refining a 50,000+ word non-fiction book covering macroeconomics, geopolitics, geoeconomics, techno-economics, and sociopolitics.

The initial architecture should include three coordinated agents: an Architect to manage the thesis, chapter plans, and persistent book-context ledger; a Researcher to retrieve and organize evidence from supplied documents and approved web/data sources; and a Writer to produce section-level drafts from approved outlines and research briefs. Include human review checkpoints before major drafting stages.

Implement RAG for custom PDFs and research material, using a practical vector-store solution such as Qdrant, Chroma, or FAISS. The workflow should support structured context/memory, source-aware research notes, prompt caching where available, and section-by-section drafting to reduce repetition and factual drift.

Deliver a documented codebase, setup and execution instructions, and a simple CLI or lightweight interface to run the pipeline, monitor stages, and export generated chapter drafts. Python and LangGraph, CrewAI, or a comparable orchestration approach are acceptable.

Budget range:
Very small project ($250 – 750 USD)

About the client
I am a Btech Engineer and CFA charterholder. I have co-authored two books on macroeconomics + geopolitics + sociopolitics. Looking to write a third one with agentic AI help.