WoCo PMS
A multi-tenant HR performance-management platform
145 → 35ms
Auth latency
4.5 → 8.6/10
Code quality
~4–5×
Throughput
Backend optimization program
Overview
WoCo PMS is a multi-tenant HR performance-management platform used by client organizations to run their entire review process — a React SPA backed by a Node/TypeScript API. It replaces slow, spreadsheet-driven annual reviews with a continuous, real-time performance system.
The problem
Traditional performance reviews are annual, manual, and disconnected — HR teams juggle spreadsheets for goals, 360° feedback, and appraisals across many client organizations, with no single source of truth and no real-time visibility.
The solution
A multi-tenant platform that runs the full review cycle in one place: goal and KPI tracking, 360° feedback with reviewer-visibility rules, per-cycle performance timelines, and an AI question generator (OpenAI) that drafts weightage-balanced KPI, 360, competency, and PLI questions. Each tenant is isolated with its own data, and a bulk Excel pipeline lets HR import questions at scale.
Architecture
A multi-tenant SaaS: three React SPAs (Employee, Manager, HR) talk to a layered Node/Express API, which serves data from MongoDB, MySQL, and Redis and calls OpenAI for AI features. High-level overview.
React SPAs
Employee · Manager · HR
Node/Express API
layered · multi-tenant
MongoDB · MySQL · Redis
data · cache
OpenAI
AI question generator
Key features
- Goal & KPI tracking across review cycles
- 360° feedback with configurable reviewer-visibility rules
- Per-cycle performance timelines and evaluation windows
- AI question generator (OpenAI gpt-4.1-mini) for KPI/KRA, 360, competency, and PLI questions
- Bulk Excel upload for question authoring at scale
- Star-rating system with scale normalization across the app
- Multi-tenant architecture with per-tenant data isolation and S3 document uploads
Tech stack
Frontend
Backend
AI
Infrastructure
My role & impact
- Core engineer and #1 frontend contributor — integrated 100+ APIs, built 200+ reusable MUI components, and implemented Redux Toolkit state management
- Led a backend refactoring & performance initiative that raised the codebase's documented quality score from 4.5 to 8.6/10
- Cut authenticated-request overhead ~145ms → ~35ms (~4× faster) with a fail-open Redis cache, and added 44 MongoDB indexes for 100–500× faster indexed queries
- Built the AI question-generator UI and the bulk Excel-upload flow, and hardened the platform — closing 10 SQL-injection points and fixing 25+ critical bugs