K Kordis Lab Blog INSIGHTS // ENGINEERING // PRODUCT

Technical knowledge for people who build serious products.

Architecture, scalability, UX, management and technology applied to the real world. Articles written by engineers who are in the field every day.

Build vs. Buy: when to build a technology in-house or buy a ready-made solution

The wrong question is "build or buy?". The right one is "is this a competitive advantage or just operations?" The decision framework that separates the two — and avoids both wasted rebuilds and vendor lock-in.

Read full article
Latest articles

Recent content for product and engineering teams.

DevOps

Kubernetes in production: 12 mistakes we see almost every month

Default configs, mis-sized requests, aggressive liveness probes and other self-inflicted wounds that get expensive in B2B environments.

Architecture

Event-driven in practice: when the extra complexity is worth it

Messaging patterns, idempotency, and the classic mistake of turning everything into an event when a synchronous call would do.

AI

Enterprise RAG: how to build semantic search that actually works

Ingestion pipeline, chunking, embeddings and re-ranking. What separates a nice-looking POC from a production-ready solution.

UX/UI

Design tokens: the invisible asset that separates amateur products from serious ones

How to structure primitives, aliases and component tokens at scale — and why Figma tooling alone doesn't solve it.

Product

MVP is not a bad product: the method we use to validate in 6 weeks

Cutoff criteria, validation metrics, and how to avoid the "MVP that became the product" trap without ever going through real discovery.

Management

How to choose a software house without regretting it by month 3

The technical, contractual and cultural checklist every CTO should apply before signing any development SOW.

Architecture

Migrations without downtime: the 4-phase pattern for changing schema in production

Renaming a column shouldn't require a maintenance window. The expand-contract pattern for changing schema in production without downtime.

UX/UI

Continuous discovery for small teams: the 4-hours-a-week framework

How to keep qualitative research running without burying the product team in interviews nobody will watch later.

DevOps

Observability that won't cost you a kidney: the open-source stack we use with clients

Prometheus, Grafana, Loki and Tempo configured with sanity. The metrics that actually matter for SLOs and when to ignore the rest.

VAMOS DAR O PRÓXIMO PASSO

Seu desafio merece uma boa conversa.

Conte como sua operação funciona e o que você gostaria de transformar.

Vamos conversar