Kamyar Kian

HCDS Architect. Agentic AI & Civic Tech.

Projects

Toronto Shelter Capacity Audit

A quantitative analysis of Toronto’s shelter system using Python & Open Data. This audit revealed critical saturation points in specific sectors (e.g., Women & Youth services hitting 99% capacity) and visualized systemic gaps to support data-driven policy advocacy.

PythonPandasMatplotlibOpen Data APICivic Tech

Cognitive State Agent (LangGraph)

An advanced Agentic AI architecture moving beyond stateless RAG. Built with LangGraph, this system maintains "Cognitive State" to handle multi-turn reasoning and crisis management, simulating a proactive teammate rather than a reactive chatbot.

LangGraphPythonLLM EngineeringState ManagementSystem Design

Wildfire Economic Predictor

A novel predictive model exploring "Economic Empathy" in climate algorithms. This project analyzes real-estate market fluctuations as leading indicators for forest fire risks, bridging the gap between financial data and environmental safety.

Scikit-LearnPredictive ModelingFinancial Data AnalysisPython

Recent Signals

Stop Building Chatbots, Start Building Teammates

2026-01-01

Why "Stateless" RAG fails in crisis scenarios—and a technical blueprint for building Cognitive State using LangGraph. A shift from reactive bots to proactive agentic workflows.

The Price of Smoke: Why Climate Algorithms Need "Economic Empathy"

2025-12-08

A technical proposal for predicting forest fires using Real-Estate Market signals instead of just thermal sensors. How economic data can act as a leading indicator for environmental disasters.