Engineering × AI × Automation × Project Systems
I build systems that turn operational problems into structured workflows, decision tools, and automation with explicit evidence, verification, and human control.
My work combines engineering reasoning, project execution, AI-assisted decision systems, and software automation.
| Project | What it demonstrates | Live |
|---|---|---|
| Opportunity Cockpit | AI-assisted analysis, evidence, uncertainty, verification, Human Gate | Demo |
| CP Engineering | Requirements, scope, engineering risk, cost, quote, revision and decision readiness | Demo |
| Wii Ops Starter | Agentic operations, proportional autonomy, verification, evidence-driven closure | Demo |
These public repositories are intentionally separated from their private operational cores. They use synthetic data and exclude credentials, personal applications, customer information, proprietary thresholds, private infrastructure, and protected technical material.
AI-assisted opportunity decision system
A system for discovering, normalizing, evaluating, and prioritizing professional opportunities without converting missing information into invented facts.
Key ideas:
- structured opportunity intake;
- evidence-aware fit analysis;
- explicit unknown states;
- independent verification;
- Human Gate before external action.
Traceable engineering project cockpit
A workflow connecting requirements, scope, risks, concept selection, internal cost, commercial quote, project revisions, and decision readiness.
Key ideas:
- engineering information provenance;
- risk as decision support rather than automatic approval;
- cost separated from sale price;
- revision-aware decisions;
- qualified human decision gate.
Agentic operations and governance system
An internal control plane for coordinating projects, agents, automation, verification, quality gates, evidence, and controlled execution.
The public starter demonstrates:
- preflight before execution;
- separation of change risk from execution authority;
- proportional autonomy;
- worker/verifier separation;
- evidence-driven closure.
A public Next.js application exploring B2B real-estate pre-sales workflows and decision support.
- Evidence before claims.
- Human gates for consequential external actions.
- Simple architecture until scale or complexity justifies more.
- Reversible changes and explicit rollback paths.
- Tests and verification before declaring work complete.
- Automation should reduce manual work without hiding operational risk.
Python · TypeScript · React / Next.js · FastAPI · Docker · GitHub Actions · SQLite · Supabase · Playwright · AI/LLM integrations
Technology is selected per problem; the portfolio is organized around systems and outcomes rather than tool lists.
Public GitHub contains selected demonstrations and applied projects. Private repositories may contain operational data, project history, infrastructure details, personal information, customer information, or proprietary implementation and remain intentionally private.
Forked repositories are kept for experimentation and reference and are not presented as original portfolio work.
- GitHub: @Flavio459
- LinkedIn: Flávio Souza Barros
Last reviewed: September 2026.