<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Vorrapong Kertnat</title><description>Full-stack developer in Bangkok writing about TypeScript, ElysiaJS, DevSecOps, and a slow pivot into financial engineering.</description><link>https://blog.icerust.dev/</link><item><title>Why is ElysiaJS fast? Inside its route-specific handlers</title><link>https://blog.icerust.dev/blog/why-is-elysiajs-so-fast/</link><guid isPermaLink="true">https://blog.icerust.dev/blog/why-is-elysiajs-so-fast/</guid><description>A look inside Sucrose and how Elysia generates request handlers, separating framework JIT, engine JIT and build-time AOT, with the trade-offs and how to evaluate the gains on a real API.</description><pubDate>Fri, 02 Oct 2026 00:00:00 GMT</pubDate></item><item><title>What is an M.Sc. in Financial Engineering? What you study, and what background helps</title><link>https://blog.icerust.dev/blog/what-is-financial-engineering/</link><guid isPermaLink="true">https://blog.icerust.dev/blog/what-is-financial-engineering/</guid><description>What Financial Engineering is, what you study, what it costs, and how to prepare — notes from the first 3–4 weeks of an M.Sc. at UTCC.</description><pubDate>Sat, 19 Sep 2026 00:00:00 GMT</pubDate></item><item><title>My deploy broke and my repository did not change</title><link>https://blog.icerust.dev/blog/wrangler-auto-config-broke-my-deploy/</link><guid isPermaLink="true">https://blog.icerust.dev/blog/wrangler-auto-config-broke-my-deploy/</guid><description>A green deploy in August, an identical pipeline in September, and a build that failed twice. The cause was a tool my CI downloads fresh on every run, answering its own setup prompt with yes.</description><pubDate>Sun, 13 Sep 2026 00:00:00 GMT</pubDate></item><item><title>Elysia vs Hono with PostgreSQL on a MacBook Air M5</title><link>https://blog.icerust.dev/blog/elysia-hono-postgres-on-m5-air/</link><guid isPermaLink="true">https://blog.icerust.dev/blog/elysia-hono-postgres-on-m5-air/</guid><description>A repeated, single-machine comparison using a shared Bun SQL database layer, with downloadable measurement evidence.</description><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate></item><item><title>The slowest framework drew the least power</title><link>https://blog.icerust.dev/blog/energy-per-request-on-m5-air/</link><guid isPermaLink="true">https://blog.icerust.dev/blog/energy-per-request-on-m5-air/</guid><description>I measured what an HTTP request costs in joules on a fanless MacBook Air M5. Express is 2.3x slower than Elysia but only 1.8x hungrier — because it cannot draw as much power. Then I found out the load generator was using four times more CPU than the server it was measuring.</description><pubDate>Wed, 26 Aug 2026 00:00:00 GMT</pubDate></item><item><title>Benchmarking Bun HTTP frameworks on a MacBook Air M5</title><link>https://blog.icerust.dev/blog/bun-http-framework-benchmark-on-m5-air/</link><guid isPermaLink="true">https://blog.icerust.dev/blog/bun-http-framework-benchmark-on-m5-air/</guid><description>Reversing the order of the target list reversed the ranking. What a fanless laptop actually measures when you benchmark eight Bun HTTP frameworks back to back, and what the numbers look like once you stop doing that.</description><pubDate>Sun, 23 Aug 2026 00:00:00 GMT</pubDate></item></channel></rss>