Founder, FX Studio AI · Finance Executive · AI Solutions Architect
I connect executive experience in Finance, Accounting, Controllership and enterprise transformation with the engineering of practical AI systems.
At FX Studio AI, I design solutions involving agents, multi-agent systems, RAG, automations, integrations and orchestration. The aim is business value that can be measured: lower operating cost, higher productivity and revenue enablement, stronger governance, and protection of margin.
Em português: Fundador da FX Studio AI, executivo de Finanças e arquiteto de soluções de IA. Uno domínio financeiro e engenharia para transformar problemas empresariais em sistemas governáveis e úteis.
- Translating finance and controllership processes into explicit rules, data contracts and decision flows.
- Designing AI architectures that combine retrieval, tools, agents, approvals, observability and human review.
- Building automations and integrations that fit existing operations instead of creating isolated experiments.
- Connecting technical choices to cost, revenue, productivity, governance and margin questions.
These repositories are the public evidence I use to explain how I work. They are technical showcases; where stated, they use synthetic data or reference architectures. They do not claim client results.
- Pipeline-Orchestrator — multi-agent execution governance with deterministic gates, adversarial review, audit-trail artifacts and fail-closed controls.
- skill-advisor — local-first routing across skills, plugins, MCP servers and agents, with bounded composition, prompt-injection sanitization and an automated Node.js test suite.
- loopify — deterministic compilation and execution of agent loops, with explicit step contracts and executable tests.
- Brazilian Tax Reform Oracle — public RAG showcase with a synthetic tax corpus, citation-oriented responses, and a runnable Python retrieval prototype.
- IFRS 15 Revenue Intelligence — synthetic FastAPI/Pydantic/SQLite prototype for performance obligations, allocation and revenue schedules.
- Controllership Reconciliation Engine — synthetic Python prototype covering aging, ABC analysis, risk classification, support gaps, top accounts and monthly trends.
Together, these projects show evidence across agent engineering, RAG, automation patterns, integrations, architecture, security boundaries, testing and governance — always tied back to a business domain.
Financial reporting · controllership · close and consolidation · planning and analysis · revenue recognition · lease accounting · tax and reconciliation domains · internal controls · operating-model transformation
Agent and multi-agent orchestration · RAG and citation flows · workflow automation · system integrations · architecture and contracts · security boundaries · deterministic tests · governance and human-in-the-loop controls
FX Studio AI is the studio through which I explore and build enterprise AI solutions. Its focus is the convergence of business context and engineering discipline: understand the process, make the rules explicit, design the control points, and ship an auditable system.
The public repositories intentionally separate technical evidence from confidential work. No client data, private metrics, certifications or employment claims are presented here unless they are explicitly documented in the repository itself.