Skip to content

ECHO AI Skunkworks

ECHO (Expert Coding for Hydrologic Observations) is a WMA-authorized AI Skunkworks team that rapidly prototypes AI-assisted tools for camera imagery and hydrologic workflows, evaluates feasibility, and hands off successful innovations to production teams for deployment.


Overview

Founded in 2026, ECHO establishes a rapid-prototyping capability to evaluate AI-assisted tools before committing production resources. The team of 5–7 developers explores applications of machine learning and large language models to camera imagery analysis, automated quality assurance, and hydrologic data workflows.

Mission

  • Rapid prototyping — evaluate AI tools quickly without production overhead
  • Feasibility assessment — determine which AI approaches are viable for operational use
  • Technology transfer — hand off successful innovations to production teams
  • Risk reduction — test emerging AI capabilities before large-scale investment

Focus Areas

  • Automated image quality assessment for NIMS imagery
  • AI-assisted water level detection from camera imagery
  • Natural language interfaces for hydrologic data queries
  • Computer vision for automated scene classification
  • ML-enhanced image velocimetry preprocessing

My Role

I proposed and secured WMA leadership authorization for ECHO, establishing the team charter, recruiting developers, and setting the technical direction. I serve as Founder and Development Team Lead.

Status

Active — FY2026–present

Funded by Next Generation Water Observing Systems (NGWOS).

v1.5.0.0

Welcome to IVyTools v1.5 🎉

This is a major update bringing interactive lens correction, smarter search line handling, improved velocity visualization, and friendlier error reporting. Here's the highlights:

Plus: improved orthorectification for portrait/oblique cameras, add/remove GCP rows directly in the table, documentation migrated to MkDocs, and numerous bug fixes for units, uncertainty, and project loading.

Tip: Existing v1.x projects will load normally. Reprocessing is recommended to benefit from the search line clipping and ortho improvements. Older projects using 3D rectification will need to be review, as the new 3D method will be slightly different (XS will move).

Full details in the v1.5.0.0 Release Notes.