The Problem (Why)

Colgate-Palmolive (a fortune-500 enterprise CPG corporation) needed an enterprise-grade solution for managing their growing portfolio of multi-tenant B2B2C microsites. The existing architecture was highly fragmented across multiple frameworks and data sources. A lack of standardized practices created operational bottlenecks, made tracing application bugs difficult, and introduced performance risks that threatened to inflate infrastructure costs.

High-level overview of B2B2C microsites architecture

The Technical Challenge and Solution (How)

We needed to unify a complex multi-tiered ecosystem that integrated identity management (Gigya/Okta), enterprise ERP (SAP S/4HANA), and external APIs (Shopify Storefront) across a B2B admin portal and DTC customer-facing ecommerce frontend.

To resolve the fragmentation, I architected and maintained a robust system:

• Transitioned the Node.js (Express), React (Next.js), and Vue (Nuxt) stack to Hexagonal Architecture by establishing a robust provider/adapter and ports pattern. I separated business logic from external services, enforcing clean code principles and enabled true multi-tenancy.
• Unified disparate data sources using Prisma (PostgreSQL) and overhauled the caching layer (Redis).
• Governed standardization across engineering teams by implementing strict Gitflow practices, Semantic Versioning (SemVer), and CI/CD enhancements to ensure reliable deployments.

The Results

The architectural shift enhanced both end-user experience and internal DX (developer experience). By replacing fragmented legacy patterns with best-in-class system optimizations, the platform achieved enterprise-level stability.

Highlights:

• Resolved a critical per-instance in-memory leak by implementing a unified Redis caching strategy. This directly increasing service uptime and eliminating sporadic service interruptions.
• Reduced overall infrastructure costs by optimizing system resource usage and streamlining the deployment pipeline.
• Enhanced traceability by re-engineering GCP logging strategy using a Winston transport layer. This give the engineering team immediate, highly accurate observability into system health and dramatically reducing lead time (time-to-resolution) for platform bugs.