Reuben Okeke Senior Go Developer
Raleigh, NC • reuben.okeke@gmail.com • +1 919-555-0148
Profile Summary
- Senior Go Developer with 9 years of experience building control plane services and internal tools for managed OpenShift and container platforms, specializing in gRPC services, concurrency and performance tuning.
- Hands-on with Go (Go 1.27, generics), APIs (gRPC, net/http), data (PostgreSQL, Apache Kafka), cloud (Kubernetes, AWS) and observability (Prometheus, OpenTelemetry).
- Deep expertise in goroutine and channel design, gRPC contract design, Kubernetes operators and pprof-based tuning, using context-first cancellation and benchmark-backed changes to ship Go services that stay fast and predictable under load.
- Works closely with SRE, Product, Security and Support teams and platform engineers at customer companies, in design reviews, on-call rotations and upgrade planning for thousands of clusters.
- Mentors engineers on idiomatic Go and concurrency reviews, runs a weekly Go code review clinic, and wrote the team’s service template and error-handling guide.
Technical Skills
- Languages:
- Go 1.27, SQL, Bash, Python, TypeScript (reading)
- Go Libraries & Frameworks:
- net/http, chi, Cobra, client-go, controller-runtime, sqlc, pgx
- APIs & Protocols:
- gRPC, Protocol Buffers, REST, OpenAPI, grpc-gateway
- Databases & Messaging:
- PostgreSQL, Apache Kafka, Redis, NATS, Amazon DynamoDB
- Cloud, Containers & CI/CD:
- Kubernetes, OpenShift, AWS, Docker, Helm, Terraform, GitHub Actions
- Testing & Code Quality:
- go test, testify, Testcontainers, native fuzzing, golangci-lint, govulncheck, benchstat
- Observability & Profiling:
- pprof, runtime/trace, Prometheus, OpenTelemetry, Grafana, slog
Education
Work Experience
- Rebuilt the cluster upgrade-scheduling service for the managed OpenShift fleet in Go 1.27 on net/http and gRPC, with Protocol Buffers contracts, per-call deadlines and graceful shutdown, cutting p99 latency from 410 ms to 84 ms.
- Rewrote the fleet health poller with errgroup worker pools, a semaphore capping calls to each cluster API at 16 and context deadlines on every probe, so all 2,400 clusters get checked every 60 seconds instead of every 10 minutes.
- Moved cluster event storage to PostgreSQL through pgx and sqlc and wrote the event consumer on Apache Kafka, with idempotency keys and a dead-letter topic, processing 5.8M events a day with no duplicate customer notifications since launch.
- Profiled the cluster inventory API with pprof heap and CPU captures, removed per-request JSON re-encoding, pooled buffers with sync.Pool and set GOMEMLIMIT per pod, cutting memory per replica from 1.9 GB to 620 MB.
- Added table-driven tests, Testcontainers suites against real PostgreSQL and native fuzzing of the cluster spec parser to CI, with the race detector on every run, cutting bugs that reached staging from 9 a quarter to 2.
- Wrote a controller-runtime operator with Kubebuilder that reconciles add-on installs across managed clusters through a custom resource, with status conditions, finalizers and envtest coverage, cutting install time from 45 minutes to 6 minutes.
- Built a Cobra CLI that lets support engineers inspect and repair stuck cluster state, shipped as signed static binaries for Linux, macOS and Windows through GoReleaser, cutting median diagnosis time from 40 minutes to 8 minutes.
- Led the move of 11 service repositories from Go 1.22 to Go 1.27, added govulncheck and Dependabot to CI and pruned unused modules, taking open vulnerable-dependency findings from 37 to 0.
- Added a Redis cache in front of Elasticsearch in the Go service behind Docker Hub image search, cutting median search latency from 320 ms to 70 ms.
- Wrote integration tests for the Hub webhooks service in Go with Docker Compose fixtures, catching 23 regressions before release in the first year.
- Replaced 4 Python scripts for image cleanup with one concurrent Go CLI, cutting the nightly run from 3 hours to 25 minutes.