About

I spent twenty years selling the future. Now I build it.

The short version: I know this technology because I've sold it, deployed it, and run it at home on my own dime.

Joseph Schafer
The professional story

From the demo room to the deployment.

I've spent two decades at the leading edge of business intelligence and process automation — Tableau, Alteryx, Databricks, and lately the agentic AI tooling that's reshaping how work actually gets done. For most of that, my job was to help the world's biggest companies understand new technology well enough to bet on it: technical discovery with CIOs and architects, proofs of concept that had to survive contact with reality, and the business cases that turned a promising demo into a funded program.

Along the way I led sales-engineering teams and, in my best year, ranked first in the company for new-business ACV. But the part I kept gravitating to wasn't the pitch — it was the build. The moment where you stop talking about what the technology could do and actually wire it up to see whether it holds. So I leaned all the way in: I now do the building myself, for businesses that need someone who can tell real innovation from hype because they've shipped both.

Today that means independent consulting on the South Shore of Massachusetts — emerging-tech evaluation, data engineering and integration, analytics that drive a decision, and helping teams actually adopt the AI tooling everyone's talking about. The throughline is the same one it's always been: make systems that were never meant to agree do exactly that, then hand someone a result they can act on.

The human version

Off the clock.

I live in Scituate, on Boston's South Shore, and I'm the kind of person who keeps a four-machine GPU cluster in the basement — partly solar-powered — so I can run local models and pressure-test new tooling on my own dime. It's how I form opinions worth paying for: if I'm going to tell you something is real, I've usually already broken it myself first.

My wife runs a specialty grocery down the street, which is how a chunk of my "consulting" turns into reconciling point-of-sale data at the kitchen table. I write at questionabledecisionmaking.com when something's worth thinking out loud about, and I build small things — a Discord bot, this website service — mostly because I can't help it.

Credentials at a glance

The resume version.

Now

Independent Data & AI Consultant

// Scituate, MA · remote
  • Emerging-tech evaluation and POCs across AI, agentic systems, and data infrastructure.
  • Data engineering & integration — pipelines, Databricks, REST APIs, reconciliation.
  • Analytics & decision support; AI tooling enablement for teams adopting agentic workflows.
  • Productized small-business websites, built and hosted end to end.
20 years

Sales Engineering & Solutions Leadership

// enterprise BI & automation
  • Ranked #1 in company for new-business ACV; 172% of quota in a recent fiscal year.
  • Led and scaled a sales-engineering team (4 → 8), running 25+ concurrent deal cycles.
  • Technical discovery and POC delivery with CIOs and enterprise architects.
Selected work

Things I've built

// proof, not slides
  • Three Sisters Provisions — POS/inventory/ordering integration on Databricks + GCP, with demand forecasting validated across 1,400+ forecast-days.
  • BrnzyBot — containerized Discord bot with a MIP solver and subscription billing, serving 30+ users.
  • The home cluster — four machines, substantial combined VRAM, partly solar, for local inference and tooling evaluation.
Toolbox
Tableau Alteryx Databricks GCP Python REST APIs Prophet Agentic AI / Claude Code Local LLMs / Ollama Cloudflare Discord / Polar