Senior Platform Engineer & DevOps Lead · 8+ Years · Kubernetes · Go · eBPF
I build platform infrastructure for high-traffic systems serving 20M+ users — custom Kubernetes operators in Go, low-level Linux/eBPF security tooling, and Internal Developer Platforms that cut cloud overhead and accelerate delivery. Open source contributor to the Terraform and Cisco ecosystems.
- ip-pool-operator — Kubernetes operator that auto-syncs remote IP/CIDR allowlists (CDNs, cloud providers) into
NetworkPolicyresources- Fetches IP ranges from external URLs on a configurable interval (default: 1h) with deduplication and
#-comment handling - Flexible namespace targeting via label selectors; supports Ingress/Egress with port/protocol control
- Multi-architecture images (linux/amd64, linux/arm64); built on Kubebuilder / controller-runtime
- Supports providers like Cloudflare, Arvan Cloud, bunny.net, and Sotoon out of the box
- Fetches IP ranges from external URLs on a configurable interval (default: 1h) with deduplication and
-
nat-ebpf-xdp — High-performance NAT (SNAT/DNAT) implemented in eBPF, attached at the XDP hook for near line-rate packet translation before the Linux kernel network stack is involved
- Full TCP, UDP, and ICMP support with connection tracking via LRU hash maps (automatic eviction of stale entries)
- Zero-copy packet processing at the XDP hook; 2–4 Mpps in generic mode, 10–20 Mpps in native driver mode
- Incremental L3/L4 checksum updates per RFC 1624; runtime-tunable port pools and per-protocol timeouts
- Ring-buffer event streaming, live connection dump, and statistics reporting for observability
-
glbp — Linux daemon (Go) implementing GLBP-inspired gateway load balancing with virtual IP election and rotating virtual MACs
- Multicast UDP hello protocol with priority-based AVG election and optional preemption
- Round-robin ARP MAC distribution to spread client connections across multiple routers
- Uses libpcap for ARP I/O; YAML-based config with configurable hello/hold timers
- terraform-provider-haproxy — Terraform provider for managing HAProxy configuration via the HAProxy Data Plane API, enabling infrastructure-as-code workflows for load balancer resources
- Manages frontends, backends, servers, ACLs, bind sockets, and global/defaults sections as Terraform resources
- Supports runtime configuration changes through the Data Plane API without requiring HAProxy restarts
- Validates configuration before applying to catch errors early in the plan phase
- Full end-to-end, unit, and integration test suite; published to the Terraform Registry
- s3-ceph-exporter — Prometheus exporter for monitoring Ceph RGW / S3
- Grafana dashboards included
- End-to-end, unit, and integration tests
- ansible-role-haproxy — Ansible role for HAProxy + Keepalived + Data Plane API
- Multi-instance HAProxy with Keepalived; idempotent execution; fully customizable
- ansible-etcd-cluster — Etcd cluster setup with or without TLS; includes a dedicated teardown role
- ansible-hadoop-hdfs — Hadoop HDFS cluster automation with disk/DNS management and safe cluster teardown
- kafka-docs — Apache Kafka reference documentation from fundamentals to advanced patterns
- Covers Kafka concepts, CLI tooling, producer/consumer implementation, and advanced topics
- Structured progressively for developers at any level
- ansible-tutorial — Step-by-step Ansible guide with practical examples and best practices
- Linkerd-Flagger-Tutorial — Linkerd & Flagger integration tutorial with troubleshooting tips