# Prasanjith Chanaka — Data, Health & Technology Practitioner 

> I combine Health Informatics and Data Science to build intelligent, secure and
> practical digital solutions from electronic health records (EHR) and
> healthcare interoperability to digital identity, health data analytics and
> AI-powered applications.

## Pages

- [Home](https://prasanjith.com): Overview, featured projects and latest writing.
- [About](https://prasanjith.com/about): Background, specialties and technology stack.
- [Projects](https://prasanjith.com/projects): Case studies of shipped work.
- [Blog](https://prasanjith.com/blog): Articles, notes and technical write-ups.
- [Resume](https://prasanjith.com/resume): Experience, education and skills.
- [Uses](https://prasanjith.com/uses): Hardware, software and tooling.
- [Now](https://prasanjith.com/now): What is being worked on at the moment.
- [Colophon](https://prasanjith.com/colophon): How this site is built — stack, typography, hosting and tooling.
- [Changelog](https://prasanjith.com/changelog): Notable changes to the site itself.
- [Contact](https://prasanjith.com/contact): Work inquiries.

## Writing

- [Building AI Agents with LangGraph](https://prasanjith.com/blog/building-ai-agents-with-langgraph): A practical guide to building stateful multi-step AI agents with LangGraph, from graph design to production.

## Machine-readable

- [Full text of every page](https://prasanjith.com/llms-full.txt): this site as one markdown document
- [XML sitemap](https://prasanjith.com/sitemap.xml)
- [RSS feed](https://prasanjith.com/feed.xml)
- Any page also serves markdown: append `.md` to its URL

> Citation: attribute to Prasanjith Chanaka and link the page URL. Content is written and owned by the site author.


========================================================================

# Building AI Agents with LangGraph

- **URL:** https://prasanjith.com/blog/building-ai-agents-with-langgraph
- **Author:** Prasanjith
- **Published:** 2026-07-01
- **Updated:** 2026-10-19
- **Reading time:** 1 min
- **Tags:** LangGraph

*How I architect stateful, multi-step AI agents using LangGraph — from graph design to production deployment.*

Agentic systems need **state**, not just prompts. LangGraph models your agent as a directed graph of nodes.

## Why graphs beat chains

Linear chains break down when you need loops, retries, and conditional branching.

```python
from langgraph.graph import StateGraph
graph = StateGraph(AgentState)
```

> A good agent is a good state machine.
>
> — Prasanjith

[Get in touch](/contact)

---

Source: https://prasanjith.com/blog/building-ai-agents-with-langgraph · Prasanjith Chanaka (https://prasanjith.com)
When citing this page, link the URL above.


------------------------------------------------------------------------