varsha joshy✳
← All work

NATIONAL FIRE CHIEFS COUNCIL / TECHNOLOGY FOR GOOD

Better insight.
Safer communities.

Architecture for an AWS-hosted machine-learning fire-risk platform.

MY CONTRIBUTIONTechnical architect · Cross-team design
PERIOD2026
TOOLKITAWS · Machine learning · Platform architecture

01 / THE CHALLENGE

The problem
worth solving.

A fire-risk model needs a usable platform around it: a way to run it, serve its outputs and help users explore the results.

02 / MY CONTRIBUTION

What I built.

As technical architect for Force for Good 2026, I defined the end-to-end architecture for the team’s machine-learning prediction and serving platform.

01

Connect the whole product

Defined how model execution, backend services and the website fit together on AWS.

02

Give teams a shared design

Provided architecture across frontend, backend and ML delivery so implementation could progress against a common plan.

03

Design for exploration

Connected prediction outputs to the user-facing fire-risk website the team built.

03 / THE OUTCOME

What changed.

An 11-person team built the platform against a shared architecture, connecting machine-learning delivery with backend services and an AWS-hosted website.

Visuals are illustrative studies of the workflow, not screenshots of employer systems. Project descriptions focus on my contribution.

NEXT CASE STUDY

Validation frameworks

→