Blog
Insights, tutorials and news from the Nordic Byte engineering team.
-
Vector databases: when your AI feature actually needs one
A vector database is not the default answer for every AI feature. Here is how to work out whether semantic search earns its operational cost,…
-
Qué gana de verdad una pyme con un software a medida (y qué no gana)
Cuatro beneficios que se cuentan en horas y en errores, no en adjetivos. Y las tres cosas que un desarrollo a medida no os va…
-
Data retention: what to keep, what to delete, and what it actually costs
A practical look at data retention as an engineering problem: database design, backups, logs and analytics, not just a compliance document.
-
Time-series database or plain PostgreSQL: where should IoT telemetry actually land?
How to decide whether IoT telemetry belongs in plain PostgreSQL, a partitioned Postgres, or a dedicated time-series database.
-
Stripe or GoCardless: the billing engineering decisions that matter more than the fees
Choosing Stripe or GoCardless is the easy part. The webhooks, state machine and dunning logic underneath decide whether your billing actually holds up.
-
Backup and disaster recovery: what to actually ask a software vendor
RPO, RTO, the 3-2-1 rule and restore testing: the concrete questions a UK business should ask before trusting a vendor with its data recovery.
-
Enterprise SSO in B2B SaaS: when to build it before a customer asks
SAML and SSO often surface mid-deal in a security review. Here is how to decide when a small B2B SaaS product needs to build it…
-
¿Tu Kit Digital se quedó corto?
Gastaste el bono, instalaste el software, y tu equipo sigue con el Excel de siempre. Diagnostícalo tú mismo en una hora.
-
Predictive maintenance from IoT data: when it is worth building
What predictive maintenance actually requires beyond a sensor feed: failure history, labelling, model choice, and the false-positive cost that decides if it is worth building.