Portfolio — knowledge architecture, AI engineering, and outdoor media by Jeremy Henricks

Open to knowledge, support and applied-AI roles

// 01

$30M+ Support cost savings

KCS / KM ARCHITECTURE

Knowledge architecture at enterprise scale

Global KCS programs through two of the largest acquisitions in tech — 160K+ articles under governance, 1,500+ publishers and 300+ coaches mentored, mid-90s CSAT held through both integrations. 20–25% case volume reduction

  • KCS
  • M&A Integration
  • SearchUnify
  • Python / NLP
Explore Enterprise Architecture
// 02

<24h From problem to live product

AI ENGINEERING

Real products, shipped at AI speed

The night before a mountain hike I was juggling four tabs — wildfire smoke, weather, traffic, pass cameras. A day later 511now.com was live: 2,700+ official cameras plus fire, weather and traffic data in one fast site. Grok Bot and other AI tools collapsed the build cycle; the judgment — sourcing rules, honest labels, curated answers — is mine. Live at 511now.com · Washington, Oregon & Idaho

  • AI-Assisted Development
  • Vanilla JavaScript
  • Leaflet
  • Official Data Feeds
Read the 511Now Story
// 03

21M Monthly platform users

OUTDOOR MEDIA & NETWORKS

Platforms that carried the outdoors online

Designed and launched OutdoorHub.com for Carbon Media Group — syndication architecture, taxonomy and SEO that carried a niche media network to 21 million monthly users, and the discipline behind the outdoor products I still build. 33% revenue growth · #1 outdoor network

  • WordPress
  • Syndication
  • Information Architecture
  • SEO
Read the OutdoorHub Case Study
All 20 projects
  1. 01

    Structure the knowledge

    KCS article structure, ownership, content health and governance — so what exists is accurate, current and findable in the first place.

  2. 02

    Make it retrievable

    Taxonomy, search tuning, stop words, metadata and information architecture. Retrieval quality is decided long before a model is involved.

  3. 03

    Let AI answer from it

    Curated answers first, then cited retrieval from real content, and a model fallback that is explicitly the last resort — not the default.

The current work

AI systems built on governed knowledge

Assistants, recommendation engines and generation tools where the interesting problem is never the model — it's what the model is allowed to say, where the answer came from, and what happens when the inputs don't cover the question.

Where it started

Content platforms and information architecture

Large-scale publishing systems, taxonomy, and the structural work that decides whether a site is findable — including one platform that reached 21 million monthly users.

Shipped independently

Products I designed, built and released

Full-stack web apps, a live mountain-conditions dashboard shipped in under a day, a cross-platform mobile app, audio signal processing, and the publishing system behind books that are actually for sale. Designed, architected and implemented by me, and labelled honestly — including the prototypes.

Everything, in one list

All 20 projects, oldest discipline to newest build.

Résumé version

Let's talk about the work

I'm looking for knowledge management, digital support, self-service and applied AI work — places where someone who has run these programs at scale and can also build the thing is useful. Leadership or hands-on; I care more about the problem than the title.