Case Studies

Case studies that show how technical decisions were actually made.

This section gathers semi-anonymous cases focused on diagnosis, trade-offs, and operational impact. It is a more direct way to show how I think and decide in real contexts.

Case Study

How I unstuck a Node.js/Kafka pipeline without rewriting the architecture

Editorial case study on batch processing refactoring in Node.js with Kafka and MongoDB: stable event loop, constant memory usage, and a surgical refactor delivered in hours.

Stable event loop during large batchesConstant memory usageRefactor delivered in hours, not weeks

Read case study

Case Study

How I cut over 80% of a SaaS platform's AWS bill without freezing the product

Case study on cost optimization for a SaaS platform running on AWS, with strategic use of Cloudflare, per-tenant cost tracking, and a cumulative reduction close to 85%, all without pausing development.

Cumulative reduction close to 85%Per-tenant cost now measurable, no longer estimatedDevelopment and new features never stopped

Read case study

Case Study

How I built this entire site on the Cloudflare edge

The engineering case behind this site: Astro on Cloudflare Workers, with D1, R2, KV, Vectorize, Workers AI, and an MCP server at the edge. Honest decisions and trade-offs.

All at the edge: Workers, D1, R2, KV, Vectorize, Workers AINo traditional server to maintainAutomatic global deploy on every push

Read case study