Vancouver, BC · Systems that pay for themselves

Demos don't ship. Dashboards don't decide. I build the systems that do.

I am an industrial engineer who spent fifteen years on the floor before writing a line of production code. Today I own master data in Dynamics 365 Business Central at a regulated producer in BC, where a data error is a regulatory event, not an ugly chart. I build the data, the analytics and the automations your operation actually runs on. Shipped, tested and working. Not a pilot.

See what I've shipped → Book a call

01 / What I do

Two services. Boring on purpose.

I don't sell «AI transformation». I sell two concrete things: data your operation can trust, and a supply chain that stops leaking money.

Business Central · Master data & analytics

Master data, BOM configuration, routings, MRP readiness and migrations in Dynamics 365 Business Central. This is the critical, tedious work partners subcontract, because it is easy to get wrong and expensive when it is. On top of it, the analytics layer: Power BI, SQL and Power Query models that turn clean data into decisions your team trusts. I do this every day inside a Health Canada regulated environment, where the data has to be right.

D365 BCMaster DataBOM / MRPSQLPower BIPower Query

From CAD 120/hr · by project or partner subcontract

Supply chain

Planning, inventory and operations analysis for plants and distributors that grew faster than their processes. Bottleneck diagnosis, inventory sizing, S&OP readiness and flow redesign, backed by an MSc in Logistics and the ISCEA certification. I turn fifteen years on the floor into models your planners actually use, not spreadsheets they quietly ignore.

PlanningInventoryS&OPFlow analysisISCEA

By project · starts with a scoped diagnostic

Working stack

Power BI SQL Python Power Query n8n

02 / Case study

Koci Lead System

A renovation contractor in Vancouver was invisible to the exact people most likely to hire him: someone who just bought a condo with a dated bathroom. So I built the machine that finds them and shows them their own bathroom, already renovated.

In production

From MLS listing to mailbox, automatically

The system watches recently sold condos in Metro Vancouver, filters those matching the ideal client, takes the real bathroom photo, re-renders it as a finished renovation preserving the real geometry (walls, windows, plumbing), builds a two-tier priced flyer and mails it physically to the new owner. Every send passes human approval.

1,002

TESTS PASSING

397

LISTINGS INGESTED

SPRINTS SHIPPED

$0.04

COST / AI RENDER

Pipeline

Ingest

Apify to realtor.ca
sold listings

Qualify

Claude classifies
usable photos

Render

Gemini rebuilds
the bathroom

Compose

WeasyPrint
2-page flyer

Approve

Human gate
before spend

Mail

PostGrid prints
and ships

spend_guard Per-service daily caps and a hard monthly ceiling. The system cannot spend the client's money without permission. That is a feature, not a limitation.

Stack

PythonFastAPISQLite + volumeJinja2WeasyPrintApifyAnthropic ClaudeGoogle GeminiPostGridRailwaypytestRBAC dashboard

03 / Selected work

A track record across stacks and industries

Rimex Invernaderos · Full ERP (ASP.NET + SQL)

An end-to-end ERP I built for a greenhouse manufacturer on ASP.NET and SQL Server. The whole company ran on it: inventory, purchasing, administration, BOM and production, plus accounts payable and receivable and greenhouse-project progress. One connected system, one version of the truth, every department on the same data.

ASP.NETSQL ServerERPBOM / ProductionInventoryAP / AR

Ask me about it →

Rubicon Organics · D365 migration

Redesigned a workflow that cut manual processing by 90%. Master data migration and BOM configuration for a Dynamics 365 implementation in a Health Canada regulated plant, including the link between production schedules and Bartender labeling with full compliance across a complex SKU matrix.

Dynamics 365Master DataBartender

Ask me about it →

n8n job-posting automation

An n8n workflow that scrapes new job postings and fires notifications on its own. The kind of small, reliable automation that quietly deletes a daily manual chore. Built and running.

n8nScrapingNotifications

Ask me about it →

Beansprouts Kidz · Daycare website

A clean, fast, findable website for Beansprouts Kidz, a local daycare. Built to load quick and get found by the parents already searching for it.

WebSEO

Ask me about it →

Quinta El Capricho

Bilingual direct-booking site for a vacation villa in Tequesquitengo, Mexico. Hand-built static, custom i18n, WhatsApp booking with no backend. Full LodgingBusiness JSON-LD validated green in Google Rich Results, plus llms.txt for answer-engine visibility.

Statici18nJSON-LD

quintaelcaprichoteques.com ↗

04 / Who

Edgar Santana

I spent fifteen years in operations before I wrote a line of production Python. That order matters. I know what a wrong BOM costs, what a stockout does to a plant, and why a floor supervisor does not care about your model architecture.

Today I own master data governance in Dynamics 365 at a licensed cannabis producer in BC, where a data error is a regulatory event, not just an ugly dashboard. On my own time I build data and automation systems for businesses that want results, not decks.

Industrial engineer. MSc in Logistics from Mondragon. MBA from University Canada West. Trilingual. Based in Burnaby, working across Canada, Mexico and remote.

05 / Contact

What's your bottleneck?

If it's data quietly breaking your ERP, or a supply chain leaking margin, those are the things I fix. Tell me the bottleneck. If I am the right person, I will show you how. If I am not, I will point you to who is.

hello@esantana.co · LinkedIn ↗