SCALABILITY
How to Scale a Node.js API from 100 to 1M Requests
Each scaling tier exists to remove one specific bottleneck. Here is which bottleneck, in what order, and what each fix costs you.
6 min read
Engineering notes, written down.
What I have learned building production systems — architecture, scale, reliability and AI pipelines. Each article states plainly whether it describes something I built, something demonstrated on this site, or a reference design.
Latest
SCALABILITYReference design
How to Scale a Node.js API from 100 to 1M Requests
Each scaling tier exists to remove one specific bottleneck. Here is which bottleneck, in what order, and what each fix costs you.
6 min read Read article
All articles
GENAI
How Production RAG Systems Actually Work
The demo is a vector search and a prompt. The production system is eight components, and most of the engineering is in the seven that are not the model.
5 min read
BACKEND
Redis Queues vs Direct API Processing
When to do the work inside the request, when to queue it, and the four things that make a queue safe to retry.
5 min read
SYSTEM DESIGN
Designing a Multi-Tenant SaaS Architecture
Three isolation models, how to pick one, and why the dangerous bug is always a query somebody forgot to scope.
4 min read
ENGINEERING
How to Handle Third-Party API Failures
Every external dependency will fail. Timeouts, breakers, backoff and a defined degraded state are the four things that decide whether that becomes your outage.
5 min read
DATABASES
What Happens When Your Database Goes Down?
Read-only mode, buffered writes and a rehearsed failover turn a database outage into degraded service instead of a dead product.
4 min read
FRONTEND
Angular SSR vs CSR for SaaS Applications
Rendering strategy is a per-route decision, not a per-app one. Marketing pages, dashboards and documentation each want something different.
4 min read
Working on something like this?
If any of these problems look like yours, the fastest way forward is a short technical conversation.