Vinicius Silva
EN PT ES
Résumé

Senior Software Engineer · Backend & Full Stack

Engineering reliable software for real-world operations.

5+ years building production enterprise systems. Backend first: Java and Spring Boot microservices, REST APIs and event-driven integrations with RabbitMQ, PostgreSQL and AWS. Full-stack delivery with React, Angular and TypeScript, and hands-on AI-assisted development with Claude Code, Codex and AI agents.

Get in touch View selected work Résumé (PDF)(opens in a new tab)

Open to full-time, international senior software engineering roles. Currently building an industrial operations platform independently, directly with the client.

Based in Blumenau, Santa Catarina, Brazil · UTC−3 · Working remotely with international teams

Engineering snapshot

5+ years
Professional software engineering
Enterprise
Product teams and regulated domains
International
English-speaking, distributed teams
End-to-end
Requirements, architecture, delivery, evolution

Selected Work

Systems built and owned end to end.

A few projects that show how I work — from the operational problem to the running system. Selected, not exhaustive.

System view

How I think about a system.

A generic, anonymised architecture: the shape of the platforms I build. No client data, only structure.

  1. People use web and mobile clients.
  2. A REST API validates input and owns the rules.
  3. State changes become events on a bus.
  4. Independent services react to those events.
  5. Data is persisted, observed and kept evolvable.
Web · Mobile REST API · Spring Boot PostgreSQL Event bus · RabbitMQ Domain services Observability Web · Mobile REST API · Spring Boot PostgreSQL Event bus · RabbitMQ Domain services Observability

Industrial manufacturing client · Mar 2026 – present

Industrial Operations Platform (MES / WMS)

Read the full case study →

Problem
A plastic-injection manufacturer tracked materials, moulds and production in spreadsheets, labels and memory. Finding stock, knowing what had happened to it and knowing what each machine was doing got slower as volume grew.
What I built
A modular, on-premise MES/WMS platform, as sole engineer. Warehouse management (delivered): locations, 2D/3D views, drag-and-drop operations, audit history, role-based access and PDF reports. Machine monitoring (pilot in validation): a Node.js edge app reads machine digital inputs and outputs every 300 ms over HTTP and Modbus-TCP, detects cycles, buffers offline in SQLite and resends; the backend ingests events idempotently, validated by sequence. Production planning: work-order lifecycle, per-machine queue, downtime and scrap tracking. Client care: versioned proposals with acceptance criteria and change control, and a documentation site with a plain-language user manual built from real screens plus a full API reference (OpenAPI specification, interactive Swagger, PDF).
Stack
Java · Spring Boot · PostgreSQL · Docker · Node.js · Modbus-TCP · SQLite · React Native / Expo · JWT
Result
Warehouse management runs on the factory network. 2,400+ automated tests across the edge app, the backend (JUnit 5 / Mockito) and the UI (Jest), written in an AI-assisted workflow with Claude Code and Codex. Planned, not built yet: the full machine fleet, MQTT, and OEE / MTBF / MTTR indicators.

Senior Sistemas · GED / ECM · digital signature

Document & Signature Platform

Problem
Document-management and digital-signature products used by thousands of corporate accounts, split across dozens of interdependent microservices that had to ship without breaking each other.
What I built
Worked across 12 components of the suite (5 Angular front ends and 7 Spring Boot services): REST APIs on PostgreSQL with document storage on AWS S3, PDF stamping and audit reports, bulk import and batch operations, and asynchronous, event-driven integrations on RabbitMQ.
Stack
Java · Spring Boot · Angular · PostgreSQL · RabbitMQ · AWS S3
Result
Features delivered end to end, releases kept free of breaking changes across services, and log traceability driven with Kibana and Logstash.

Cast Group · Studio Z · e-commerce

E-commerce Restocking Platform

Problem
An e-commerce restocking platform that depended on external systems, and where slow data retrieval was limiting the operation.
What I built
Java / Spring Boot microservices with SAP and SOAP / WSDL integrations, deployed on Red Hat OpenShift, in a Scrum team. Later defined and led enhancement packages, turning client requirements into deliverable scope.
Stack
Java · Spring Boot · Microservices · SAP · SOAP / WSDL · SQL · OpenShift
Result
Complex SQL queries engineered and optimised, increasing query speed by 40% and reducing database response time.

Cast Group · enterprise and public sector

Multi-client Engineering

Problem
Enterprise and public-sector teams that needed full-stack software: manual drafting of government procurement notices, a legacy JSF logistics system, statewide apps.
What I built
Full-stack delivery for an international industrial group (a project run entirely in English), a state court, a steelmaker and a state IT company, with React, Angular and Vue.js front ends over Java, Spring Boot, Quarkus and Node.js services.
Stack
Java · Spring Boot · Quarkus · Node.js · React · Angular · Vue.js
Result
A browser extension that cut days of drafting to minutes (state court), a legacy JSF-to-Quarkus migration (steelmaker), and statewide school asset-tracking and hospital scheduling apps.

Client-specific operational, financial and infrastructure detail is omitted or anonymised.

Portfolio

Real projects first, then product suites.

A local multi-agent system with its own MCP server, an industrial MES/WMS platform and an AI-assisted engineering method, followed by three product suites you can open in the browser with fictional data.

Browse the portfolio

AI-Assisted Engineering

AI as an engineering tool, with the engineering kept in.

I use Claude Code and Codex every day, and I build agents, skills and an MCP server of my own. The rule stays the same: nothing counts as done without a test or other evidence.

  1. Daily workflow

    Claude Code and Codex from design through automated testing. On the industrial platform this workflow produced 2,400+ automated tests: 80+ on the edge app, 490+ on the backend with JUnit 5 and Mockito, and 1,900+ on the UI with Jest.

  2. Agents and MCP

    Hive is a local, hierarchical multi-agent system (master, group leaders, specialists) with a Go backend and a React dashboard. Its MCP server over stdio lets Claude Code or Copilot ask the agents and search my notes.

  3. Context and retrieval

    Agents get their context from hybrid retrieval over an Obsidian vault: keyword search combined with embeddings generated by a local server. Notes enter prompts as untrusted text.

  4. Method and skills

    PAIRED: Problem, Analysis, Implementation, Review, Explain and Document, with decision gates and a fast and a production mode. Plus custom Claude skills for orchestration and review.

What I do not claim: production experience with RAG at scale or with vector databases. The retrieval above runs on my own projects.

See Hive and PAIRED in the portfolio →

Experience

Where I've done this.

Mar 2026 – Present · Independent Software Engineer (Contractor)

Industrial Operations Platform

Industrial Manufacturing Client · Santa Catarina, Brazil

Designing and building a modular, on-premise MES/WMS platform end to end for a plastic-injection manufacturer: Java / Spring Boot, PostgreSQL, React Native / Expo (web and mobile) and Docker, running on the factory network. Warehouse management is delivered; real-time machine monitoring is a pilot in validation; production planning covers work orders, per-machine queues, downtime and scrap. Built with an AI-assisted workflow (Claude Code and Codex) and 2,400+ automated tests. Senior engineer and sole owner, not a company or an agency.

Read the case study →

Employment history

Senior Sistemas

Software Developer IV

Sep 2025 – Jul 2026

Document management (GED/ECM) and digital-signature products at one of Brazil's largest ERP vendors, used by thousands of corporate accounts. The suite runs on dozens of interdependent microservices; I worked across 12 of its components (5 Angular front ends and 7 Spring Boot services).

Delivered features end to end: backend REST APIs on PostgreSQL with document storage on AWS S3, plus the Angular screens for the document vault and the signing portal. Implemented PDF generation, stamp placement and audit summaries with PDFBox, iText and JasperReports, and contact management with bulk import and batch operations. Designed asynchronous, event-driven integrations on RabbitMQ and managed cross-service version compatibility to keep releases free of breaking changes. Drove log traceability with Kibana / Logstash, led staging and production troubleshooting, and enforced code reviews and JUnit / Mockito tests in CI/CD. For a period, as Squad Leader, ran release planning, scope and estimations across backend and frontend.

  • Java
  • Spring Boot
  • REST APIs
  • Angular
  • PostgreSQL
  • RabbitMQ
  • AWS S3
  • PDFBox / iText / JasperReports
  • Kibana / Logstash
  • JUnit / Mockito
  • CI/CD

Cast Group

Full-Stack Engineer (Trainee → Senior Development Analyst)

Aug 2021 – Sep 2025

IT consultancy delivering full-stack software for enterprise and public-sector clients: Studio Z (e-commerce), Bosch (an international project run entirely in English), TJMG (the Minas Gerais Court of Justice), Usiminas and Prodemge. Grew from trainee to senior development analyst across the engagement.

Studio Z, e-commerce (2021–2024): built Java / Spring Boot microservices with SAP and SOAP / WSDL integrations and Red Hat OpenShift deployments in a Scrum team; engineered and optimised complex SQL queries for the restocking platform, increasing query speed by 40%; defined and led enhancement packages. Bosch (2022): requirements, refinements, estimations and sprint reviews in English, delivering with TypeScript, Node.js, Next.js and MongoDB. Prodemge (2023): statewide school asset-tracking and hospital scheduling applications with Java and Vue.js. Usiminas (2024–25): migrated legacy JSF logistics systems to Java / Quarkus services with Angular front ends and CI/CD on Jenkins and GitLab. TJMG (2025): founding team of a browser extension that automates government procurement notices, with a React front end and Spring Boot APIs, cutting drafting from days to minutes; managed the CI/CD lifecycle with Jenkins and Red Hat OpenShift across development, staging and production.

  • Java
  • Spring Boot
  • Quarkus
  • React / Next.js
  • Angular
  • Vue.js
  • TypeScript
  • Node.js
  • SQL
  • MongoDB
  • SAP · SOAP / WSDL
  • Jenkins / GitLab
  • Red Hat OpenShift

Full history on LinkedIn(opens in a new tab) Download résumé(opens in a new tab)

Expertise

What I bring to a system.

Grouped by capability. Filled marker: core, used daily. Hollow: working knowledge.

  1. Backend & Distributed Systems

    • Java (Core)
    • Spring Boot (Core)
    • REST APIs (Core)
    • Microservices (Core)
    • RabbitMQ (Core)
    • Event-driven architecture
    • Node.js
    • Quarkus
    • SOAP / WSDL
    • JWT / RBAC

  2. AI-Assisted Engineering

    • Claude Code (Core)
    • Codex (Core)
    • AI agents (Core)
    • MCP (Model Context Protocol)
    • Claude skills
    • Prompt engineering
    • Local LLMs (llama.cpp)
    • Retrieval over embeddings

  3. Data

    • PostgreSQL (Core)
    • SQL (Core)
    • Spring Data JPA / Hibernate
    • Flyway
    • MongoDB
    • SQLite

  4. Full-Stack

    • React (Core)
    • Angular (Core)
    • TypeScript (Core)
    • Next.js
    • Vue.js
    • React Native / Expo
    • RxJS

  5. Cloud & Delivery

    • AWS (Core)
    • Docker (Core)
    • CI/CD (Core)
    • Git (Core)
    • Jenkins
    • GitLab
    • Azure DevOps
    • Red Hat OpenShift (Kubernetes)
    • Elasticsearch / Logstash / Kibana

Engineering Practice

  • JUnit / Mockito (Core)
  • Code review (Core)
  • System design (Core)
  • Agile / Scrum (Core)
  • OpenAPI / Swagger
  • Release planning
  • Jira

Filled: core. Hollow: working knowledge.

Proof

Where to check my work.

Career history and code GitHub(opens in a new tab) LinkedIn(opens in a new tab)

How I Work

One loop, whether it's a team's codebase or a client's problem.

  1. 01

    Understand

    Start from the operational problem and the people who have it, not the feature list.

  2. 02

    Design

    Choose the simplest architecture that satisfies the real requirements, and make the trade-offs explicit.

  3. 03

    Build

    Implement in small, verifiable increments, with tests that prove behaviour.

  4. 04

    Validate

    Check against the requirements and real usage; fix what the feedback surfaces.

  5. 05

    Evolve

    Keep the system maintainable as it grows, and treat the roadmap as hypotheses, not promises.

About

About

I'm a Senior Software Engineer with 5+ years building software for enterprise product teams and, more recently, directly for an industrial client. My focus is backend engineering in Java and Spring Boot — REST APIs, event-driven microservices, distributed systems — with the full-stack range in React and Angular to deliver a complete product. I work with AI tools every day (Claude Code, Codex and agents and skills I build myself), and I own the work from the first conversation to production.

Right now I'm building an industrial operations platform end to end for a manufacturing client — in the room for requirements, architecture, trade-offs and the long tail of maintenance and evolution. That end-to-end responsibility is the part of the work I care most about; it's an engineering engagement, not a company or an agency.

Location
Blumenau, Santa Catarina, Brazil · UTC−3
Work
Remote, with international teams
Languages
Portuguese (native) · English (advanced, C1)
Vinicius Silva — Senior Software Engineer · Backend & Full Stack

If any of this matches what you're hiring for or building, the next step is a short conversation.

Get in touch

Contact

Let's talk.

Open to full-time, international senior software engineering roles. Currently building an industrial operations platform independently, directly with the client.

I reply within 48 business hours.

Get in touch LinkedIn(opens in a new tab) GitHub(opens in a new tab) RésuméPDF · EN (opens in a new tab)