I design and run low-latency, high-throughput backends in Go—event-driven services that stay fast, observable, and cost-efficient at scale. WebSockets infrastructure for hundreds of thousands of concurrent viewers and Kafka→ClickHouse analytics that ingest ~1M events/day.
My focus: predictable throughput, low tail latency, cost-efficiency, and operability under load. I favor simple, measurable designs (clear SLIs/SLOs, backpressure, idempotency) and invest in observability, so failures are shallow and cheap.
What I bring:
- Throughput & reliability: event-driven Go services with bounded queues, circuit-breakers and observation instrumentation.
- Data at speed: Kafka / Redpanda (partitioning & keys that match access patterns), ClickHouse (aggregated views, TTLs, rollups).
- Ops rigor: Kubernetes, GitOps CI/CD, Terraform & Ansible for infrastructure as code (IaC), alerting that reflects user impact.
- Cost discipline: lean storage/egress, right-sizing, and aggregation strategies that cut waste.
Core stack: Go
, MongoDB
, Kafka
, Redpanda
, ClickHouse
, Redis
, WebSockets
, Kubernetes
, GitOps
, Prometheus/Grafana
.
Industries: live video, collaborative streaming, usage analytics, SaaS backends.
Open: Remote (EU). Relocation: France.