The longer version.

I came to software through economics and client work. At PUCPR in Curitiba, I studied incentives, risk and the costs behind business decisions. Working with businesses showed me how often those costs came from fragmented information, manual handoffs and processes nobody could reliably repeat.

I wanted to change those processes. In April 2020, I began building software for clients, working across interfaces, backend services, databases, integrations and deployment. The work usually began with a conversation about a slow operation rather than a finished specification.

That became my consulting practice. At Cliniq, I built Python pipelines and dashboards to consolidate operational data and expose capacity bottlenecks. E-commerce engagements involved order and inventory integrations, where reconciliation, retries and partial failures mattered as much as the initial API connection.

Client work taught me to ask who owns the next step, what happens when an external service fails, and how someone else will operate the software after handover. Those questions still shape my architecture decisions.

In 2023, I built SkytOffer, a product for diagnosing gaps in commercial offers. Users supplied their audience, promise and offer structure; the software organized that context into recommendations about positioning and buyer objections.

Building a product changed the problem. I could explain missing context during a consulting call, but software needed to handle incomplete inputs and make its next step understandable without me in the room. That experience became the foundation for AlturionX.

From November 2024 to February 2026, I ran AlturionX as its technical founder and software engineer. I owned architecture and implementation across Python/FastAPI services, Next.js/TypeScript interfaces, PostgreSQL, cloud deployment and AI-assisted operational workflows.

One engineering problem was duplicated research: separate sections of an analysis made their own external requests and could work from inconsistent sources. I consolidated those requests into one shared evidence bundle per analysis. Model outputs passed schema validation before persistence, and conclusions referenced the evidence used to produce them.

I also coordinated collaborators across engineering, commercial and delivery work. Client conversations influenced implementation priorities, but each custom request still needed a decision about maintenance and whether it belonged in the shared product.

Consulting continued alongside AlturionX. From November 2024 to January 2026, I developed Node.js services on AWS for Inova Recurso, bringing client data, funding criteria and document rules into reusable workflows with human approval controls. At Metabio, I built Python reporting workflows combining campaign data, behavioral signals and AI-assisted interpretation.

In February 2026, I completed a private asset sale covering AlturionX's codebase, operating platforms and client portfolio. I prepared transition documentation and a structured handover. Making the work understandable to its next owner was part of the engineering responsibility.

My public projects explore related engineering questions. CareLoop covers tenant-scoped workflows and queue-backed processing. Fencier checks coding-agent diffs against repository policy. AtlasOS separates numerical computation from AI narration and validates reported figures against stored artifacts. These are public engineering projects, distinct from the client engagements above.

Outside work, Curitiba is home. I am engaged, and most late afternoons I walk around Parque Barigui with Mick, my small white dog and least impressed critic. The walk gives me time away from the screen and a useful boundary between work and the rest of my life.

I am looking for a senior software engineering role focused on backend, AI systems or platform work. I bring six years of client and product engineering, hands-on implementation, and experience taking responsibility from the first requirements conversation through deployment and handover.