Enterprise Live Data Sync Framework - Real-Time Log Processing Pipeline
Engineered a 25-command Redis-compatible server from scratch in C++ implementing the RESP wire protocol over raw TCP, fixing a payload-truncation bug and processing 13,672 logs across 3,070 lines of code with zero data loss. Developed a two-stage validation and delivery pipeline (Python/FastAPI) achieving a 94.3% cloud acceptance rate via token-bucket rate limiting and exponential backoff with jitter, cutting retry storms across 2,758 requests despite a 5.7% injected fault rate. Orchestrated 7 Dockerized services with a multi-stage build, exposing Prometheus metrics scraped every 5s across 3 targets to a 6-panel Grafana dashboard for real-time pipeline observability.
// INTERVIEW ENGINE