Back to professional work
Production Internal Platform · Upay · 2026

Upay KPI Workspace

A production internal platform that turns multi-role KPI operations into one controlled, auditable workflow, from monthly target setup and submissions through review, correction, and reporting.

[ EVD ]

Engineering evidence

My contribution
End-to-end engineering ownership · Discovery, architecture, implementation, hardening, and rollout
Outcome
Deployed to production for internal KPI operations
Decision record 01
Kept authorisation and workflow transitions server-side, and ran PostgreSQL in CI as in production, so concurrency and integrity are tested against the real database engine.
[ ARC ]

Architecture

Described at pattern level only: the system is confidential, so no internal data, endpoints, infrastructure, or business rules are shown.

  1. 01

    Web client

    • Next.js + TypeScript
    • Role-specific views
    • Excel import preview / export

  2. 02

    API

    • Django REST Framework
    • Per-action capability checks
    • Guarded workflow transitions

  3. 03

    Domain & audit

    • Transactional state changes
    • Append-only audit history
    • Bounded file handling

  4. 04

    Data & ops

    • PostgreSQL (CI and production)
    • Redis
    • Docker, health checks, backups
[ KEY ]

Technical highlights

  1. 01

    Modelled a multi-stage monthly operating lifecycle as guarded state transitions instead of loosely connected forms and manual handoffs.

  2. 02

    Enforced capability-based authorisation and record-level confidentiality on the server, with regression tests for access, elevation, and cross-user data boundaries.

  3. 03

    Built the production path around versioned containers, PostgreSQL-aligned CI, health checks, backup and rollback procedures, audit integrity, and bounded file handling.

[ ABS ]

Abstract

Upay KPI Workspace covers the complete lifecycle of KPI operations across several departments. My work spans workflow discovery, product boundaries, full-stack architecture, authorisation, transactional domain logic, testing, deployment, and operational readiness. Because the system is confidential, this case study describes only engineering patterns and outcomes; it publishes no internal data, infrastructure addresses, proprietary rules, source repository, or production access path.

[ STK ]

Stack

  • Next.js
  • TypeScript
  • Django REST Framework
  • PostgreSQL
  • Redis
  • Docker