Skip to content

Full-Stack Developer

VINCENT JAMES SALVINO

Specializing in backend systems, API design, and database optimization. Experience building production internal platforms, role-based access control systems, real-time applications, and workflow automation tools used across multiple departments. Skilled in React, Next.js, Node.js, Laravel, PostgreSQL, and MySQL. Passionate about scalable architecture, developer tooling, and AI-assisted software systems.

View projectsExperience
GitHubLinkedInPili, Camarines Sur, Philippines

Selected projects

3 projects shown.

01

Personal Project

  • Shared, persistent memory for coding agents — an MCP server any client (Claude Code, Codex, Cursor) can connect to against the same SQLite database.
  • Fuses memory recall, code/dependency indexing, and ranking into one token-budgeted context pack per task, instead of dumping raw search hits into the agent's context.
  • Ships a read-only local visualizer (memory graph, inspector, context trace) for seeing what's stored and why it was or wasn't retrieved.
  • React
  • TypeScript
  • Node.js
  • Supabase
  • SQLite
  • MCP
Source
02

Work Project

  • Built an end-to-end ticketing system for internal staff and approved contractor companies: create, route, schedule, work, resolve, and close, with status tracking, activity history, and recurring tickets.
  • Implemented role-based visibility across staff, branch Assignors, Admins, and contractor company roles, with AMS-credential login for employees and approval-gated mobile or Viber OTP login for contractors.
  • Added admin-configurable priority deadlines with due and overdue warnings, Viber and live portal notifications, and performance reports (compliance, efficiency, accuracy) with Excel export.
  • Node.js
  • MySQL
  • React
  • TypeScript
  • Viber messaging
Live site
03

Work Project

  • Architected and developed a production task and project management platform used by FarmStation staff for task tracking, collaboration, file management, and notifications.
  • Implemented a three-level permission system (SUPER_ADMIN, ADMIN, EMPLOYEE) with department-scoped authorization controls.
  • Designed a dual-service architecture where PHP managed authentication while Node.js provided business logic APIs secured through service-to-service verification.
  • Built real-time collaboration features using Socket.IO for instant updates and notifications.
  • Authored OpenAPI documentation and contract tests to improve API reliability and maintainability.
  • Node.js
  • Express.js
  • PHP
  • MySQL
  • Socket.IO
Live site

Experience

  1. IT Programmer at Farmstation

    January 2026 - Present

    Pili, Camarines Sur, Philippines

    • Developed and maintained internal workflow and management systems used across multiple departments using React, Next.js, Node.js, and Laravel.
    • Designed and implemented REST APIs with role-based access control, enforcing permissions based on department, user role, and resource ownership.
    • Improved database performance by identifying and eliminating N+1 query bottlenecks in PostgreSQL and MySQL through query optimization and eager loading strategies.
    • Implemented WebSocket-based real-time notifications and task updates, replacing polling mechanisms and reducing unnecessary server requests.
    • Managed application deployments and release workflows through Git-based CI/CD pipelines and cloud hosting platforms.
    • Contributed to architecture decisions, backend design, and system scalability for business-critical internal tools.
    Next.jsNode.jsReactPHPExpress.jsMySQLTypeScriptTailwind CSSGitJavaScriptHTMLCSSPM2REST APIs

Education

Bachelor of Science in Computer Science (BSCS)

August 2021 - July 2025

Camarines Sur Polytechnic Colleges · Nabua, Camarines Sur, Philippines

Thesis: Realtime Local Weather Station Using XGBoost

  • Built an IoT weather station with a live Flask + Chart.js dashboard tracking temperature, humidity, heat index, and wind speed in real time.
  • Trained per-feature XGBoost binary classifiers on a collected sensor dataset (~383 samples); best model (Heat Index) reached 89% accuracy, 89% precision, 89% recall, and 0.89 F1 — outperforming a majority-class baseline by ~14 percentage points.

Stack

Databases & Tools

MySQLPostgreSQLGitGitHubVercelRailwayPM2Socket.IOSupabase

Frontend

HTMLCSSJavaScriptReactNext.jsTailwind CSSTypeScript

Backend

Node.jsExpress.jsLaravelREST APIsGraphQLNestJSPHP