Steve Ngepi

- Developed and implemented a comprehensive backend testing strategy (unit, integration, end-to-end) using Jest and WebDriverIO, achieving 90–95% coverage. Integrated it into CI/CD pipelines, reducing production bugs by ~30% and speeding up release cycles. - Configured Dockerized database environments for integration and E2E testing—ensuring isolated, repeatable, and reliable test runs. - Architected a Kafka-based microservice for asynchronous email delivery, achieving 99%+ reliability under heavy load and reducing delivery latency by ~25% through SendGrid and HubSpot. - Migrated analytics infrastructure from InfluxDB to MongoDB, cutting infrastructure costs and boosting analytical query performance by ~25% via optimized indexing and query tuning. - Refactored the metadata-driven response assembly engine (used in Ask YOGI’s backend for generating astrological insights) by restructuring logic and applying caching/memoization. This cut average response build time by ~20% under peak loads—resulting in faster personalized horoscope delivery. - Enhanced backend query performance with precomputed indices and conditional joins, reducing query times by ~15% without increasing infrastructure load. - Built and maintained scalable real-time backend systems to deliver personalized astrological forecasts, ensuring responsiveness, stability, and accuracy as user demand and data complexity grew.