Software Engineer
Adsyzygy · Aug 2024 – Present
- Reduced Go CRM API p95 latency by 81% (1.6s to 300ms) at approximately 2M requests/day through profiling, PostgreSQL query optimization, connection pooling, and Redis caching.
- Increased Java/Spring Boot payment and loan-processing capacity 5x (400 to 2,000 transactions/minute) and reduced timeout-related failures 63% (8.5% to 3.1%) using Kafka-based asynchronous processing, idempotency, retries, backpressure, and transactional state management.
- Developed horizontally scalable Node.js WebSocket services with Redis and Kafka for cross-instance messaging and presence synchronization, supporting 25K concurrent connections without sticky sessions.
- Increased deployment frequency 6x (2 to 12 releases/month) by standardizing Docker, Kubernetes, Terraform, GitHub Actions, and automated deployment workflows.
- Reduced mean time to recovery (MTTR) 72% (125 to 35 minutes) through OpenTelemetry distributed tracing, Prometheus metrics, Grafana dashboards, actionable alerts, and incident runbooks.
- Improved React/Next.js initial page-load time 73% (4.1s to 1.1s) and reduced redundant API traffic 46% through server-side rendering, route-level code splitting, request deduplication, and caching.
- Built FastAPI-based LLM workflows with retrieval, tool calling, structured outputs, and validation guardrails.