RalphJavenBeso

1. Feb 2025 — May 2026

   Product Developer · Project Manager · Scrum Master · ouvar®

   Wore three hats end-to-end — ran sprint planning, daily standups, and backlog grooming as Scrum Master; owned roadmap, ticket flow, and stakeholder updates in Jira and Confluence as Project Manager; and translated business requirements into shipped product as Product Developer, owning UI design in Figma and the daily loop with developers and QA. Designed and built a full-stack project management SaaS (React, Node.js, Express, Prisma, PostgreSQL) deployed to production.

   ReactNode.jsPostgreSQLFigmaJiraConfluenceScrum
2. Sept 2024 — Feb 2025

   Platform Support · ouvar®

   Provided second-line technical support for an internal logistics platform. Diagnosed user issues, reproduced bugs and escalated to engineering with full repro steps, and maintained internal documentation and operational runbooks.

   JiraConfluenceFreshchat
3. May 2024 — Sept 2024

   Onboarding & QA · ouvar®

   Onboarded new clients onto the ouvar® logistics platform with structured training and documentation. Performed QA across platform features prior to release; logged and verified defect fixes.

   JiraConfluence
4. Mar 2023 — Sept 2023

   Microsoft PowerApps Developer · Hacktiv Colab Inc

   Built business applications on the Microsoft Power Platform for enterprise clients. Designed PowerApps interfaces and Power Automate workflows integrated with SharePoint, Dataverse, and Office 365.

   PowerAppsPower AutomateSharePointDataverse

I take a messy business problem and turn it into a working product — using AI as a multiplier, not a substitute for thinking.

1. 01

   Understand the problem

   Talk to the people who'll actually use it. Read what already exists. Write down what success looks like in plain language — before any code. Most software fails here.

2. 02

   Design the solution

   Sketch the data model first, then the UI. Decide what's in scope and what's deliberately not. Boring, well-understood tech that works beats clever tech that doesn't.

3. 03

   Use AI to write code

   AI is a productivity multiplier, not a replacement for engineering judgement. I prompt deliberately, review every line, and rewrite what doesn't fit the design. The architecture and the decisions stay mine.

4. 04

   Test & connect parts

   Wire things together early. Click through every flow end-to-end. Find what breaks before users do — then write down the cases worth keeping.

5. 05

   Think about edge cases

   What happens with no data? With wrong data? With a slow connection? Who else needs this to work, and how? The product isn't finished until the unhappy paths are.

Real problems, the approach I took, and the outcome.

Featured

2026Full-Stack SaaS

Case study

Linden — Project Tracker

Problem

An Australian logistics technology company needed a project management platform that no off-the-shelf SaaS could fit — 12 distinct user roles, three project types, and a hierarchical company → segment → brand scoping model that required custom multi-tenancy from the ground up.

Approach

Designed a multi-tenant Postgres schema across ~18 entities with company → segment → brand isolation, then built a React + Express + Prisma SaaS from scratch. Implemented role-based access control with permission flags, Gantt scheduling with task dependencies, time logging, comments with @mentions, a notification system, and parallel Kanban / Gantt / List / Calendar views. Used AI tooling as a build multiplier, owning every architectural decision and reviewing every line.

Outcome

Shipped a production-ready SaaS replacement for the planned ClickUp rollout in ~390 hours over 26 days. 92 UAT cases documented. Deployed to production on Vercel + Railway with separate staging environments. Case-study screenshots use anonymized branding (Linden) out of respect for the original client.

Screenshots· click to enlarge

Login — split-panel hero with the Linden serif wordmark and leaf markOperations Dashboard — KPIs, project pipeline by stage, risks & blockers panelDashboard — dark mode (full theming system, not a single design)Gantt — milestones, sub-tasks, dependency arrows, multi-scale timelineGantt — dark mode (same dependency logic, full theme parity)Kanban — five-column workflow with project cards and inline priority cuesProject detail — properties sidebar, tabbed sections, comments, full activity logMy Tasks — personal view with daily summary, stat cards, and time-logged progressUser Management — RBAC with 12 roles and 6 permission flags per user

SaaSMulti-tenantReact 19Node.jsExpressPrismaPostgreSQLJWTDockerVercelRailway