supervisor-pattern

A complete LangGraph guide for PhD students and researchers. From zero to multi-agent systems with Python examples (chatbots, ReAct agents, supervisor patterns, research pipelines).

Prism-inspired multi-agent orchestration framework built on LangGraph. Security-first, provider-agnostic, composable.

Multi-agent web harvesting on LangGraph's Supervisor pattern: 9 specialist agents, deterministic-first pipeline where the LLM is an enhancer, never a dependency — run-level circuit breaker, batched selector rescue, golden-set eval, MCP + FastAPI + distributed scheduler.

A FastAPI application that runs a finance analysis agent to analyze stock tickers using a streaming agent runner and various data tools (yfinance, Pinecone, sentiment analysis, etc.).

Several LangGraph graphs behind one chat stream — an LLM routes each turn to free chat or a structured flow, then hands control back. FastAPI + SSE, local model, no build step.

LangGraph supervisor with parallel specialist agents. 1.99x speedup vs sequential multi-agent, reproducible in 6s with no API key. Planner, retrievers, analyzers, verifiers, synthesizer.

Production-grade multi-agent AI system with LangGraph supervisor pattern, RAG knowledge base (ChromaDB), and Anthropic Claude — routes queries between specialized triage, scheduling, and eligibility agents

Supervisor-orchestrated multi-agent content pipeline (LangGraph) with 4-way LLM fallback (Gemini/Groq/OpenRouter/Cerebras), bounded revision loop, and a Streamlit UI.

Production-grade hierarchical multi-agent system with LangGraph — parallel specialist agents, HITL approval gates, safety pipeline, and fraud risk assessment workloads

Rahnuma is a source-grounded Agentic AI admissions advisor for Pakistani students, built with LangGraph, FastAPI, pgvector, and Next.js. It combines cited prospectus retrieval, deterministic university merit calculations, persistent student profiles, live web fallback, and human-confirmed profile updates

Capstone: Multi-Agent AI Platform orchestrating specialized agents for complex task solving