Senior Software Engineer, Platform at Northwind Labs
Senior Platform role · Remote (US) · Distributed systems team · 5+ years of Go or Rust · Production Kubernetes at meaningful scale.
Senior software engineer · Distributed systems
maya.okafor@example.dev · linkedin.example/mayaokafor · github.example/mokafor
Summary
Senior engineer with nine years building high-throughput backend systems in Go, with one production Rust service live for the past two. Most comfortable in the part of the stack between a queue and a database; happiest when the latency budget is tight and the failure modes are honest about themselves.
Experience
Senior software engineer
Harbor & Hatch — Remote · Apr 2021 – PresentOwned the redesign of the order-fulfillment pipeline from a per-tenant cron to a single shared worker pool. Throughput rose from 320 to 1,100 orders per minute at p99 under 800 ms, with no infra spend increase. Cut the on-call page rate in half by replacing three separate retry layers with one bounded retry, idempotency keys, and a documented replay path.
Software engineer
Cartograph Inc. — Brooklyn, NY · Aug 2017 – Mar 2021Shipped the geospatial indexing service in Go that the rest of the stack read from, replacing a Postgres + PostGIS path that timed out past 40k rows. Ran the migration as a five-week shadow read, then a one-week dual write, then a cutover that we never noticed on either side.
Software engineer
Larkspur Analytics — Boston, MA · Jul 2015 – Jul 2017Owned the MySQL → PostgreSQL migration of the reporting system. Wrote the per-tenant data audit that drove the cutover order, ran the dual-write window, and ended with a clean rollback path that we never had to use.
Selected projects
openpgp-batch
Maintainer of a small open-source library for batch-signing files with age-encrypted detached signatures. About 1.4k stars, used in two CI pipelines I do not own.
rfcgarden
A personal indexing tool that watches the IETF mail archive and surfaces the hundred most recently edited drafts; nothing public, mostly a way to keep my own reading habits honest.
Skills
Go (daily) · Rust (one production service, two side projects) · PostgreSQL · Redis · Kafka · Kubernetes · Terraform · Linux · gRPC · Prometheus / OpenTelemetry · Production on-call rotation habit
Dear Northwind hiring team,
Your platform posting calls out three things I have spent the last several years doing: the part of the stack that runs on a tight latency budget, the part where failure modes need to be named out loud, and the part where two teams need to share a queue without one of them quietly starving the other. I would like to do more of that work, on a team that has been writing about it long enough to be tired of the answer being "add another retry."
The bulk of my last four years has been in exactly that shape of system. The biggest was the fulfillment rewrite at Harbor & Hatch — a single worker pool replacing per-tenant crons — which took us from somewhere we could not reason about under load to a place where we ship with confidence on a Friday afternoon. I did most of that work alongside a small platform team; the design survived an engineer leaving, an SLO tightening that almost nobody noticed, and one genuine outage I will be happy to walk through.
Two specific things about the Northwind posting caught my eye. The first is the way you describe the on-call rotation — the part about "you will own the system you ship, including its 3 a.m." is the line I have written into my own job descriptions for the past three years. The second is that you are explicit about Kubernetes at scale rather than "cloud-native experience." I have spent the better part of a year debugging a control-plane-vs-data-plane race that landed in a Kubernetes upgrade, and I would welcome more of that shape of problem.
I have built my career on boring systems that are honest about themselves. Northwind is the same kind of shop. I would love to talk.
Sincerely, Maya Okafor