All Projects
Agentic AI
LangGraph
Multi-Agent
LLM
Research

Multi-Agent Research System

LangGraph-based autonomous multi-agent research pipeline

Complex research tasks require planning, web research, analysis, writing, and quality review — sequential steps that a single LLM call handles poorly. Orchestrated multiple specialised AI agents that collaborate on a research task and self-correct until the output meets a quality threshold.

January 1, 2025
9 views

Tech Stack

LangGraphLangChainChromaDBGroq APIFastAPIStreamlitPython

Overview

Built a LangGraph StateGraph with five agents (Planner, Researcher, Analyst, Writer, Critic); implemented conditional retry edges with a max-iteration guard to prevent infinite loops; added ChromaDB semantic long-term memory to avoid redundant searches; sandboxed a Python code executor for data tasks. Delivered an autonomous research pipeline producing quality-scored structured reports with cumulative memory, served via FastAPI with a Streamlit dashboard showing live agent status and scores.

The Problem

Complex research tasks require planning, web research, analysis, writing, and quality review — sequential steps that a single LLM call handles poorly and inconsistently.

The Solution

LangGraph StateGraph with five specialised agents (Planner, Researcher, Analyst, Writer, Critic) with conditional retry edges, long-term ChromaDB memory, and sandboxed code execution.
Multi-Agent Research System | Personal Website