Available full-time for the right thing
Boulder, Colorado Solutions Engineer Est. since the dial-up days

Hi, I'm
Katrina.
Marketer + analyst
who ships code.

Fifteen years of turning messy real-world problems into software and data — at startups that needed someone to wear seven hats and a couple of bigger companies that taught me why the small ones matter more. Open to full-time roles where the GIS, R, and dashboard toolkit meets work that matters.

Read the story → Get in touch
scroll, friend
01 Currently

The now page.

On the side

Founding advisor at Upstream AI

I'm the founding technical advisor (part-time) on an early-stage platform for small water and wastewater utilities — the 1 to 5 person teams who keep drinking water flowing in towns of a few hundred people. They've been running on spreadsheets and tribal knowledge for decades; we're building them something better.

Two pilot utilities live in Colorado. Pre-revenue, deliberately sized so I have bandwidth for a full-time analyst seat elsewhere — which is what I'm looking for now. → getupstream.ai

Also true three fun facts about me ↓
  • 01Fun fact

    Working alongside AI, on purpose

    Claude Code is a co-worker now. Sub-agents for testing, doc QA, pilot readiness. Wild times.

  • 02Fun fact

    Raising the generation that'll date robots

    Trying to make sure they grow up curious, skeptical, and unafraid of either humans or machines.

  • 03Fun fact

    Out on the trails

    PMBIA-certified mountain bike coach, ski seasons, and a constantly recalibrating relationship with elevation. See my coaching page →

02 The story so far

One job, in many costumes.

Every job I've ever had has been the same job in a different costume: take something messy in the real world and turn it into something useful for the people stuck inside it. Maps for the National Park Service. Mortgage emails for actual homeowners. Mountain-bike riders learning to send drops without sending themselves to the ER. Three costumes, one job. Geography was where I started — a discipline that lets you ask "what's happening" and "why is it happening here" in the same breath. Out of college I went into nonprofits using software for good, until I noticed "humanitarian help" sometimes meant exporting American capitalism in a development-aid trench coat. So I left.

That kicked off a decade of chasing product-market fit at startups. Homebot — where I figured out how to make millions of homeowners actually open a mortgage email. Teren — where I built marketing for a drone-imagery company and got to keep my geospatial brain alive. One success, one failure, and a permanent conviction that PMF is mostly stubbornness with a measurement plan. After that I bet on customer data platforms — the third-party cookie was dying, the post-cookie internet needed plumbing. Turned out it was a race to the bottom against a hundred competitors. Then Google didn't even kill the cookie. Weh weh.

I needed to recalibrate, so I got PMBIA-certified to coach mountain biking. Spent a couple seasons teaching women and youth how to get shreddy — and how to take care of their bikes, because MTB is fundamentally an equipment sport. Helped a beloved local ski + bike shop trade retail margin for service and community. Then I got stung by a bee in Fernie, BC, went into full anaphylaxis on a trail, and didn't die. Shortly after, I had a kid. Both reset the "what matters" axis in a hurry.

Now I'm raising a small human who'll grow up dating robots, and in the cracks I've gone all-in on AI for the unsexiest, most consequential vertical I can find: small drinking-water and wastewater utilities. They run on spreadsheets and tribal knowledge from the 1970s, and the climate is no longer being patient. So I co-founded getupstream.ai as a side project — building the software these operators need for a more volatile century.

03 The roles

Where I've worked.

2023 — Now
Retail Business Analyst (Sunlight Mountain Ski & Bike)
Mountain shop · Data analysis Open source

A proper data science project for a beloved local Glenwood Springs shop: cleaning messy POS data, restructuring an inconsistent product taxonomy, EDA on inventory and pricing distributions, with a roadmap toward demand forecasting and pricing optimization. Same skill set as enterprise, smaller scale, faster feedback loop. Repo on GitHub ↗

2022 — 23
Sr. Technical Consultant (Lytics)
Customer Data Platform

Got enterprise clients up and running on a CDP — translating "we want to know our customers" into queries, segments, and data pipelines that actually shipped.

2022
Marketing Automation Manager (Zayo Group)
Fiber Infrastructure

Built lead scoring, process automation, and the tech stack connective tissue. Helped lift lead conversion 15% and tied inbound to closed-won across the funnel.

2018 — 21
Head of Marketing Technology (Teren)
Geospatial Risk Intelligence Startup · Seed→A

Three years at a geospatial startup — built the web tech, the campaigns, and the persona-driven funnels that 4×'d qualified leads. Got a front-row seat to how engineering and go-to-market actually fuse.

2015 — 18
Solutions Engineering Manager (Homebot)
PropTech Startup Early team

Early at Homebot — built an email-driven product that helped real-estate agents and lenders engage homeowners, designed the data schema for ingesting messy real estate data, and shipped dashboards that made mortgage finance feel less like a black box (principal vs. interest, refi savings, all the unsexy-but-life-changing numbers). Watched the company grow up.

2012 — 15
Frontend Developer (National Park Service)
GIS & Web Mapping

Made maps that millions of park visitors actually used — including cartography work that became NPS Park Tiles, the official basemap for park websites. Built custom dashboards and web tools for NPS staff on top of ArcGIS, Mapbox, and Carto. Where I first learned that good software is mostly about respecting the user's time.

2010 — 11
Honors Thesis, Geography (Colgate University)
GIS · Spatial Statistics Published research

A Network of Alternative Economies — an exploratory spatial study of all 1,650 WWOOF host farms in the continental U.S. Regression models in ArcGIS against unemployment, natural-amenity, racial-diversity, and Bohemian indices; twelve field interviews ground-truthing the patterns in upstate New York. First time I got hooked on translating messy real-world data into something that helps a community see itself. Read the thesis ↗

2009 — 10
Research Assistant, Water Rates (Colgate University)
Political Science · Water Policy Published research

Worked with Prof. Manuel Teodoro on early research into professional mobility and water-rate structures — the seed of his peer-reviewed paper on why some utilities adopt conservation-oriented water rates and others don't. The finding: executives who arrive from outside an agency are far more likely to adopt conservation rates than those promoted within. Read the paper ↗

04 Selected work

Things I've shipped, written, or am tinkering with.

Rural water
Live product · 2024–now

Upstream AI

An operations platform for rural water utilities. Document repository with RAG-powered Q&A, AI-assisted daily reports, compliance prep that respects the human in the loop.

getupstream.ai
WWOOF host distribution map
Honors thesis · 2011

A Network of Alternative Economies

GIS + spatial-statistics study of 1,650 WWOOF host farms across the continental U.S. Regression against amenity, demographic, and Bohemian indices in ArcGIS; twelve field interviews ground-truthing the spatial patterns. Honors in Geography at Colgate.

180 pp · PDF
Published research · 2010

Water Conservation Rates Research

Undergrad research with Prof. Manuel Teodoro on professional mobility and water-rate structures — the seed of his peer-reviewed JPART paper on why some utilities adopt conservation-oriented rates. Career path predicts adoption, even controlling for climate.

JPART · 20(2):437
Cartography · 2012–15

NPS Park Tiles

Cartography and frontend work that became the official map basemap for National Park Service websites — the layer behind millions of park visitor trip-plans.

nps.gov/maps
MBA capstone · 2021

Elevated Van Journeys

Full 37-page business plan for an electric-van wellness tourism company: market sizing, Blue Ocean positioning, financials, ops plan, and the five-day Grand Circle itinerary. Co-authored with a therapist and a van-converting engineer. Pairs with the interactive route map at mappingkat.github.io/elevated-vans.

37 pp · PDF
Sunlight inventory analysis
Data science · 2023–now

Sunlight Mtn Retail Analysis

Inventory and pricing optimization for a small mountain-town ski & bike shop. Data cleanup, taxonomy restructuring, EDA on inventory + pricing distributions, with a roadmap toward demand forecasting. Jupyter notebooks on GitHub.

github.com/MappingKat
Open source · 2018

LearnOSM (HOT OSM)

Contributed to the official training material for the Humanitarian OpenStreetMap Team — guides that walk new volunteers through tracing buildings and roads so disaster-response teams get the maps they need. Translated into 23+ languages.

learnosm.org
05 Get in touch

Let's troubleshoot
something together.

I'm actively looking for a senior full-time analyst, GIS, or solutions role where the toolkit can be applied at scale. Also up for a conversation with founders building for small or overlooked operators, fellow geographers-gone-software, and anyone thinking carefully about AI in regulated, human-in-the-loop work.