01
Case study
Aug 3, 2026
We shipped a real, billing-paying SaaS in a weekend. Not a demo with a fake checkout — a product people paid for. This is what we built, hour by hour, and the honest limits of what 48 hours actually buys you.
02
Engineering
Jul 30, 2026
A million requests a day is about twelve requests a second — the number is a milestone, not a crisis. What actually keeps an API healthy at that volume is boring: the database, the queue, the cache, and an honest p99. Here is our playbook.
03
Case study
Jul 26, 2026
The story of a two-person startup that went from a Heroku bill that was eating the business to a stable platform at 100k users. The real numbers, the real mistakes, and the five lessons we would give any founder.
04
Engineering
Jul 24, 2026
Sub-second deployments are not about building faster. They are about never rebuilding, never re-pulling, and swapping the container once it proves it is healthy. Here is exactly how we make the switchover land in under a second.
05
Infrastructure
Jul 18, 2026
Cold starts are not random noise — they are a budget you can spend on the right things. Here is how we measured, attacked, and mostly eliminated cold-start latency on our serverless workloads.
06
Architecture
Jul 10, 2026
Edge functions and region-based serverless look similar on a feature matrix and behave nothing alike under load. Here is a practical decision procedure for choosing between them, and the hybrid pattern that works best.
07
Platform
Jul 2, 2026
Deploying the same artifact to five regions without breaking production took us from fragile shell scripts to a pipeline that can roll itself back. Here is how we built it, what we wired together, and what we would do differently.