Skip to content
Fabricio Duarte

Portfolio

Selected work

Data engineering, analytics, and automation in production, and 25+ products or systems in use. ECOM, SIGCL, Cocoma, Skadia, and custom work.

Product · data · AI

SIGAG — Integrated farm and livestock system

SIGAG: command center, GIS, and field assistant.

A farm app that works without signal: weighings, animal ID, body condition, and indicators (daily gain, heat stress, stocking). When the network is back it syncs and feeds dashboards. Several installations have been sold. Alerts run on n8n.

Role · Co-founder · product and data. Skadia, 2022–present.

  • Expo
  • TypeScript
  • WatermelonDB
  • Computer vision
  • LLM
  • n8n
  • Supabase

Analytics · ECOM

Institutional dashboards — Apache Superset & Power BI

Apache Superset: filters and institutional dashboard.

ECOM’s production BI engine: datasets, metrics, Mapbox, and cluster deploy. Apache Superset and Power BI dashboards, sliced per agency, embedded in the working system, auto-refreshed. Institutional use, not a lab.

Role · Data engineering. 2023–present.

  • Apache Superset
  • Power BI
  • RLS
  • SQL
  • Docker
  • Mapbox

Product · training · commercialized

SIGCL — Integrated Curricular and Legal Management System

SIGCL: institute dashboard on desktop and mobile.

Training-institute operations: campuses, enrollment, QR attendance, tuition, official minutes and transcripts in PDF with public verification, virtual classroom. Sold four times (USD 4,500 + maintenance).

Role · Product and development.

  • Laravel
  • Livewire
  • PostgreSQL
  • DomPDF

Product · estimating · commercialized

Cocoma — software bid desk

Cocoma: desktop sign-in and master rate card on mobile.

Estimates and documents software project cost: role rate card, COCOMO I and II, function points, WBS, P80 risk, and a PDF/Excel bid pack. It does not manage tasks. Sold for USD 1,500.

Role · Product and development.

  • Laravel
  • Livewire
  • Tailwind

From scattered data to a faithful metric, published and usable.

Book 25 min