Mad Archer

I build payment-focused microservices in Go, turning complex distributed transaction flows into dependable systems. I designed saga-flows without 2PC using compensation and idempotency to prevent duplicate and retry-related side effects.

I implement production-grade data and integration patterns: transactional outbox feeding Kafka and ClickHouse for post-release analytics, while keeping the “hot path” fast. I also hardened unstable external PSP interactions with timeout budgets, retry with backoff and jitter, circuit breaker, and DLQ plus safe intermediate statuses.

I’m big on measurable reliability and visibility. With OpenTelemetry, Prometheus, Grafana, and Jaeger (end-to-end trace_id and SLO alerts), I reduced payment incident MTTR and used profiling and caching to improve p95 latency.

Earlier roles strengthened my backend foundation: webhook HMAC validation, idempotent payment event handling, and state machines with audit trail and outbox. I also delivered backend services with PostgreSQL audit/idempotency design, RabbitMQ workers with retry/backoff and DLQ, and shadow traffic migrations to safely reduce endpoint latency.