StoryCareer Editorial
Published Aug 3, 2026 · 1 min read
What started as a weekend API for a side project grew into a service handling over a million requests a day — and almost melted down twice along the way.
I built a small REST API for a hobby app during a quiet period at work. A year later that API was in production at my company, and I was responsible for keeping it fast and available while the team around it kept growing.
The database was the first bottleneck: slow queries piled up during peak traffic and brought the whole service down. Adding indexes felt like magic but the real problem was N+1 queries hidden in code written months earlier.
I turned on query logging, identified the worst offenders, and rewrote the hot paths to fetch data in bulk. I added an in-memory cache for expensive lookups and put rate limiting in front of the service so a single bad client could not take it down.
Mornings are for reading metrics and investigating slow requests. Afternoons are split between writing tests for a new endpoint and reviewing pull requests from teammates who joined the service recently.
Measure before you optimize. A slow endpoint is usually slow because of one query, not because the whole stack is wrong. Learn to read the logs and the database query plan before touching the architecture.
Self-taught through a coding bootcamp, I landed my first frontend job after a year of building small projects and one very honest cover letter.
After eight years in corporate jobs, I left to build a small software product. It failed twice before I found something people were willing to pay for.
I started a weekly newsletter about marketing experiments. Two years later it pays a meaningful part of my income — and it taught me more than any job.