How we got here.
Looked like a Tableau gig in 2015. Underneath, it was already fullstack engineering with AI awareness baked in. A decade later, those instincts ship as our own products.
- 2015
Started up — Tableau and fullstack
First gig was Nielsen. Tableau dashboards, server admin, analyst training — and from day one, fullstack work ran alongside it. The two sides never really separated.
- 2017
Expanded into data engineering
The recurring pattern: fix the warehouse, fix the dashboards. We added pipeline, modeling, and ETL work to the menu. dbt adoption starts here.
- 2018
Started building Canopys
After shipping the same scheduler-shaped tool inside multiple engagements, we started building Canopys as a standalone product — recurring jobs, chained dependencies, failure alerting. Also started diagramming out ET1, the visual ETL framework we'd hand-roll into every pipeline job for years before productizing it.
- 2019
ExxonMobil onsite engagement
Multi-year onsite engagement at ExxonMobil — streaming environmental and property-safety data at scale, full-stack engineering leadership, executive dashboards across 80+ campuses. The engagement that defined how we run large enterprise programs.
- 2020
Already remote — busiest run yet
The world scrambled to go remote; we already were. The pandemic became our busiest stretch ever — engagements with Lever.co, Spotify, DiDi, Colgate, Hershey, Driscoll’s, StarKist, and more. Austin HQ stays for collab days and field work; day-to-day work is async, documented, and globally timezone-aware.
- 2021
Migrated to fullstack custom apps
SaaS is slowly decaying — bloated, locked-in, priced to seat counts no one signed up for. Folks wanted their own apps on their own infra, doing exactly what they need. We leaned into the React/Node/TypeScript side and made it the main thing.
- 2024
Trilex AI: From team-builder to internal-only
We started Trilex in 2023 as the first multi-agent team builder — letting teams compose agents by character and script, shipped as a public product. Early feedback revealed the gap: users expected it to solve everything, not just compose teams. We learned people wanted frontier models to do the heavy lifting instead. The market proved us right later: every major LLM now ships multi-agent orchestration. Rather than compete on commoditized tooling, we pivoted Trilex to closed IP — business-context-aware autonomous agents with human checkpoints, kept internal to protect our methods from easy replication. The product lives on, but privately.
- 2025
Launched ET1
Released ET1 — the first full end-to-end analytics platform built by a data consultancy while actively consulting clients. While the industry partnered with SaaS vendors and others became portfolio shops trading referrals, we built our own. Tableau was a stepping stone. We found our own clients, owned our own stack, and delivered visual DAG ETL for async pipelines, transforms, joins, and production data outputs. We productized the pattern we had been hand-rolling inside engagements for years, making us the first analytics consultancy to actually build an analytics product.
- 2026
Yee
Released Yee — the only code editor built in Austin by a data consultancy. Built without Stanford/MIT network privilege, without VCs at the door, without offshore pressure to scale cheap. Built by people doing real client work, solving real problems, treating security like a feature because mistakes cost us. Offline-first LLM coding inside compliance boundaries. USA-built. Your team trusts who built their tools.
- 2026
gato.to, Relay, and Bee
gato.to for collaborative custom-platform delivery. Relay for API-first team messaging you self-host that replaces Slack (99% cost reduction, integrations free). Bee — an AI agent team for SEO automation trained under founder Tyler Garrett's direction to learn digital marketing patterns and compound domain ranking improvements over time.
Want to be in the next chapter?
We're always scoping the next interesting engagement.