Projects
ChatGPT_Chatbot
This project implements a functional memory-enabled chatbot using LangChain, OpenAI’s GPT-4o-mini, and LangGraph’s agent orchestration. It demonstrates how agents can reason and act in multiple steps, maintain session-based memory, and integrate with external tools for live information retrieval.
The chatbot runs entirely in a Jupyter Notebook and supports both single-turn and multi-turn conversations with tool use, showcasing how LangGraph can structure conversational logic around external API access and memory state.
⸻
Features
- LangGraph Agent Executor: Manages control flow and decision-making using a ReAct-style loop.
- Tavily Search Tool: Allows the agent to fetch live web results during conversation.
- Threaded Memory: Uses
MemorySaverto persist user interactions across chat turns. - Tool Use History: Supports step-by-step breakdowns of tool calls and thought processes.
- Live Streaming UX: Streams LLM output in real time for responsiveness.
- Session Control: Each chat is scoped to a
thread_idto preserve or isolate context.
⸻
Key Insight
The true power of conversational AI comes from embedding reasoning, retrieval, and memory into structured workflows. This chatbot is more than a UI. It’s a stateful autonomous agent, built with tools and traceability from the ground up.