about ›

Built the way we build.

CoreBackend is a deliberately small backend engineering firm — led by its founder, with senior engineers engaged per project and sized to each engagement — based in Los Angeles, working remote-first. This page is how the firm itself is engineered: the decisions we run on, what each one costs us, and who you'll actually work with.

core@backend — about
$ core about
teamfounder-led · senior engineers per project · no juniors
baseLos Angeles, CA · remote-first
stackgo · postgres · redis · aws · kubernetes · terraform
billing$0 upfront · first invoice at ~50–70% delivered
$
principles ›

Decisions We Run On

Five decisions about how the firm works, written as diffs against the industry default. The red line is what we replaced. Each decision carries its cost — a principle that costs nothing isn't a principle.

decisions/01–senior-only-staffing.md proof › a named architect on every project →
-Sell the architect in the pitch, staff the project with juniors at blended rates.
+Every engineer has 10+ years in production. The person who designs your system builds it.
#cost › at capacity we turn down work — we don't scale by hiring down
decisions/02–boring-technology.md proof › built with › on all three systems →
-Résumé-driven architecture — five microservices before product-market fit.
+Go, PostgreSQL, Redis, Kubernetes, Terraform, proven queues — chosen for the team that operates the system, not the deck that sells it.
#cost › we lose the bids that want a buzzword inventory
decisions/03–labeled-claims.md proof › “modeled, benchmark-gated” →
-“10× faster” — no baseline, no denominator.
+Every figure is measured, or labeled modeled and gated by benchmarks before we treat it as real.
#cost › our numbers sometimes look smaller than competitors' — they're also checkable
decisions/04–nothing-upfront.md proof › the gates →
-A deposit before the first line of code.
+The Review is free, the blueprint is approved before build — and the first invoice comes only after roughly 50–70% of the agreed scope is delivered and working.
#cost › we carry the early risk on every engagement
decisions/05–replaceable-by-design.md proof › the exit gate →
-Lock-in dressed up as partnership — proprietary glue, a maintenance contract you can't leave.
+100% of the IP is yours. Runbooks and knowledge transfer are deliverables, and the blueprint is yours to build with anyone.
#cost › no lock-in revenue — clients stay because it works, or leave with the keys
people ›

Who You'll Work With

One architect, accountable by name, from whiteboard to production. You meet them at the Architecture Review; they stay through launch. Behind them, senior engineers — and nobody else.

the usual route › account manager pm relay ticket wall nothing ships ‹ working software · every two weeks slack · same business day › you your team architect named · accountable handoffs › none mid-project integration performance reliability senior engineers · one team your system healthy
# nothing you send us disappears — it comes back as working software
overlap › US hours + a daily window for European teams

Names and records appear here per engagement — you're introduced to your architect at the Review, not to a wall of headshots.