Advanced AI Development

AI Agent Frameworks

I build autonomous agents that can research, reason, and execute multi-step tasks. Here's my experience with the major frameworks.

Frameworks I Work With

LangChain

My go-to for RAG pipelines and document Q&A systems. Well-documented, huge ecosystem.

LangGraph

For complex workflows with cycles and state. I use it when agents need to iterate and refine.

CrewAI

Multi-agent teams where specialists collaborate. Great for research and content workflows.

AutoGen

Microsoft's framework for conversational agents. Good for code generation tasks.

Agents I've Built

Research Agent

Searches ArXiv, summarizes papers, and generates weekly digests of ML research relevant to my work.

Code Review Agent

Reviews PRs, checks for common issues, suggests improvements. Uses LangGraph for iterative refinement.

Documentation Agent

Reads codebases and generates documentation. CrewAI with researcher + writer agents.

Data Pipeline Agent

Monitors data quality, alerts on issues, and can auto-fix common problems.

Design Patterns I Use

ReAct (reason + act)Plan-and-ExecuteMulti-agent collaborationReflexion (self-critique)Tool-augmented generation

Technology Stack

LangChainLangGraphCrewAIAutoGenTavilyChromaDBPineconeFastAPI

Expertise by Sumit Chatterjee

Industrial Light & Magic, Sydney

Back to Portfolio