CloudQuarks Marketplace
Redington Group · Hexalytics
Leadership / Role: Backend Engineer (at Hexalytics)
Executive Summary
Developed scalable backend architecture for an enterprise B2B/B2C cloud marketplace facilitating vendor onboarding, catalog management, multi-tier subscription billing, and order workflows across 200,000+ store-specific SKUs.
Problem & Objectives
Enterprise clients encountered severe query latency and timeouts (up to ~40 seconds) when performing complex multi-facet filtering across 200,000+ store-specific product variants, vendor availability flags, and dynamic discount matrices.
Key Technical Contributions
- Restructured PostgreSQL query architecture, eliminating redundant join cascades and full table scans on high-cardinality catalog tables.
- Designed and deployed an AWS ElastiCache (Redis) caching tier with tag-based cache invalidation, ensuring instant updates upon catalog or pricing modifications.
- Re-engineered pagination and filtering algorithms from offset-based pagination to keyset (cursor-based) pagination, guaranteeing constant O(1) query time.
- Delivered modular Medusa.js commerce services and RESTful APIs supporting complex enterprise subscription lifecycles, vendor onboarding, and order fulfillment.
- Collaborated directly with enterprise client stakeholders, achieving 99.9% uptime during peak procurement sales events.
System Architecture & Design
Diagnosed sequential database scans and join overheads using PostgreSQL EXPLAIN (ANALYZE, BUFFERS). Engineered composite B-Tree and GIN indexes targeting JSONB metadata attributes. Built a multi-tier caching layer in Redis with deterministic cache key hashing and tag-based pub/sub invalidation, integrated into a modular Medusa.js backend deployed on AWS.