CLOSE ×AboutCareerStackWorkProcessContact
2026 · E-COMMERCE · BACKEND · DEVX AI LABS

Mayave — Real-time diamond inventory, zero data lag.

YEAR
2026
COMPANY
Devx AI Labs
ROLE
SDE-1 · Backend Engineer
Medusa.jsPostgreSQLOpenSearchStrapiNode.jsRedis
Diamond product node with ERP data pipeline streams
<1s
ERP Sync Latency
100ms
Search Response
Queue
Driven Ingestion
PROBLEM

Mayave's diamond e-commerce platform had a critical disconnect between their ERP system and the storefront — inventory updates took hours to reflect, product data was inconsistent, and the search experience was slow and inaccurate. The existing backend couldn't handle the throughput of real-time diamond catalog updates.

APPROACH

Built a production-grade backend using Medusa.js and PostgreSQL with a queue-driven data ingestion pipeline that processes ERP updates in real-time. Implemented Redis-backed job queues to handle high-throughput inventory events without overwhelming the database. Integrated OpenSearch with custom analyzers for diamond-specific faceted search — cut, carat, color, clarity — with sub-100ms response times. Connected Strapi CMS for editorial content management and SEO metadata delivery. Designed for zero data lag between ERP and storefront.

RESULT

Real-time ERP-to-storefront data sync with sub-second latency. Diamond catalog fully searchable via OpenSearch with faceted filtering. Strapi-driven content and SEO live in production. Backend handling high-throughput inventory events reliably via queue-driven architecture.

— VISUALS
BACKEND ARCHITECTURE — Queue-driven ERP ingestion pipeline with Medusa.js and PostgreSQL
BACKEND ARCHITECTURE — Queue-driven ERP ingestion pipeline with Medusa.js and PostgreSQL
SEARCH LAYER — OpenSearch with diamond-specific faceted filtering and custom analyzers
SEARCH LAYER — OpenSearch with diamond-specific faceted filtering and custom analyzers