Skip to content

Technologies

The stack, and why it is here.

Tools chosen for reliability and fit, not fashion. Every item below is in production; none of it is aspirational.

  1. 01Languages & Frameworks5 tools
  2. 02Data & Storage3 tools
  3. 03Cloud & Infrastructure3 tools
  4. 04Developer Tooling4 tools

  1. 01 / Languages & Frameworks

    Languages with a deep hiring pool, frameworks with real documentation.

    • TypeScript

      Static typing across every service and frontend we run.

    • React

      Component-driven UI, server components first where the framework allows it.

    • Next.js

      Full-stack React: routing, server rendering, and static generation in one framework.

    • Go

      Backend services on Echo, for throughput and a binary that starts instantly.

    • Tailwind CSS

      Design tokens as utilities, so the system is enforced at the markup rather than in review.

  2. 02 / Data & Storage

    Data is the most expensive thing to replace. Choose proven.

    • PostgreSQL

      The system of record, reached through pgx rather than an ORM where the query matters.

    • Redis

      Caching, rate limits, and the fan-out behind real-time updates.

    • Amazon S3

      Object storage for uploaded media and backups, addressed through the AWS SDK.

  3. 03 / Cloud & Infrastructure

    Mainstream platforms with real documentation, not infrastructure only we can operate.

    • AWS

      Managed infrastructure, chosen so your next hire has already operated it.

    • Docker

      One image per service, so local, CI, and production run the same build.

    • Vercel

      Deployment for the Next.js frontends, in the region the readers are in.

  4. 04 / Developer Tooling

    Open tools with a hiring market, not a black box we built ourselves.

    • GitHub Actions

      CI/CD in the same repository as the code, so the pipeline transfers with it.

    • OpenTelemetry

      Vendor-neutral traces and metrics, so the instrumentation outlives the backend it reports to.

    • Prometheus

      Metrics exposed by the service itself, connection pool saturation first. No dashboard stack to operate.

    • Sentry

      Errors with the stack trace attached, and no personal data on the event.

Why these, not those.

Every tool below is in production somewhere. None of it is aspirational, and none of it is here because it is trendy. We pick what the next engineer already knows, what the failure modes are documented for, and what can be handed off without a knowledge transfer project.

If a project needs a different tool, we use it. The stack is not a religion. It is a set of defaults that make the handover boring, which is the point.

Need a different stack?

Tell us what your systems already run on. We will say plainly whether it is something we work in.

contact@latellu.com
Book a consultation