https://huisam.github.io/huisam's blog — backend & AI tech engineering notes. 2026-08-12T21:18:58+09:00 huisam https://huisam.github.io/ Jekyll © 2026 huisam https://huisam.github.io/assets/img/favicons/favicon.ico https://huisam.github.io/assets/img/favicons/favicon-96x96.png 2026-08-12T19:00:00+09:00 2026-08-12T19:00:00+09:00 https://huisam.github.io/posts/llm-api/ huisam

In Coming Last time we walked through how an LLM basically works and what happens under the hood. Today I want to take a more practical angle and show how you can actually put a Frontier Model to work. The term Frontier Model may not be obvious at first, so let’s split the LLMs we commonly use into two categories: Frontier Model: trained on an enormous amount of data with a huge number o...

2026-08-10T19:00:00+09:00 2026-08-10T19:00:00+09:00 https://huisam.github.io/posts/llm-architecture/ huisam

In Coming Today I want to walk through the structure and the way an LLM (Large Language Model) works, in a way that is easy to follow. The deeper you go, the more you end up explaining mathematical formulas and paper-level material, so I am not going to dive to that depth here. Instead let’s unpack it step by step in plain language 😁 Architecture An LLM is, at its core, a model that predic...

2026-08-06T19:00:00+09:00 2026-08-06T19:00:00+09:00 https://huisam.github.io/posts/gradle-multi-module-build/ huisam

In Coming Most projects these days are structured with multi modules in mind. Teams split their code into multiple modules to separate each module’s responsibility and to push reusability further. But the moment you split into multi modules, the biggest pain point that shows up is having the same build configuration duplicated across every module. To solve exactly that, Gradle offers module...

2026-08-05T18:00:00+09:00 2026-08-05T18:00:00+09:00 https://huisam.github.io/posts/socket-exception-connection-reset/ huisam

In Coming As MSA has become the norm, we end up operating a variety of domain servers in order to work through complex business requirements. The more domain servers there are, the more network communication grows along with them. And with that growth comes a whole family of network-related exceptions. But analyzing a network error and reasoning about all the situations that could have produ...

2026-07-31T19:00:00+09:00 2026-07-31T19:00:00+09:00 https://huisam.github.io/posts/observability-spring-boot/ huisam

In Coming In our previous post, we covered the fundamental concepts of Observability. To briefly recap: Observability aims to understand why errors occur, and refers to the act of instrumentation — enabling even someone unfamiliar with a system’s internals to find the root cause of a problem. The three most important pillars of Observability are: Metric Trace Log Therefore, when o...

2026-07-29T20:10:00+09:00 2026-07-29T20:10:00+09:00 https://huisam.github.io/posts/observability/ huisam

Background Today, let’s talk about Observability. Monitoring Most of you have probably heard of the term “Monitoring.” The reason teams build monitoring systems is to check the state of their systems in real time. Monitoring is tooling or a technical solution that allows teams to watch and understand the state of their systems. Monitoring is based on gathering predefined sets of metrics ...

2026-07-28T20:07:00+09:00 2026-07-29T20:07:59+09:00 https://huisam.github.io/posts/spring-idempotency/ huisam

Introduction As systems grow larger, communication between different backend services becomes more frequent and more complex — and with more network calls comes a higher chance of network failures. These failures can stem from many causes: brief network blips, client-side connection timeouts, packet loss, and so on. Digging into the root cause every single time is expensive, and even when you...