Muhammad Saad

Reduced monthly AWS spend by 84% ($6,400 to under $1,000) by eliminating MSK and ElastiCache, consolidating 7+ microservices into a Dockerized monorepo, and right-sizing compute. Shipped an agentic AI procurement system (OpenAI API + Tavily) and architected multi-currency, timezone-aware expansion with currency-isolated ledgers.

Scaled flagship REST APIs from 3,500 to 21,000 concurrent users and improved response time from 1,200ms to 130ms (89%) using PostgreSQL materialized views, indexing, Redis caching, and connection pool tuning. Designed peak-load architecture sustaining 3,300 requests/second for flash-sale windows and led delivery of 6 revenue-critical features with a 4-engineer team.

Refactored the order placement API using SOLID principles and the Strategy Pattern, reducing latency from 6,000ms to 760ms (87%) and increasing unit test coverage from 0% to 68% (Jest + Supertest). Optimized search with Elasticsearch (custom analyzers, synonym, fuzzy matching) to reduce PostgreSQL read load by ~35% and improve checkout conversion by 5%.

Built the first production microservice (NPS module) with NestJS and Apache Kafka, scaling to 10,000 requests in under 2 minutes and establishing a reusable Kafka wrapper as an org-wide standard. Migrated high-volume events to MongoDB (notifications, OTPs, surveys) and introduced Kafka-based asynchronous communication across 4 services to reduce cascading failure risk during peak order windows.

Improved API performance by 15% by adding composite database indexes and optimizing queries on top-10 most-called endpoints. Integrated UBL as a partner bank and onboarded 3 merchants to the Checkout platform using OAuth 2.0 and webhooks.

Bachelor of Engineering in Computer Systems Engineering at NED University of Engineering and Technology (2021).