Why This Blog Exists: A Version History for Technical Growth

I want a record of learning, engineering judgement, and long-term thinking that can be revisited and tested.

阅读中文版 →
On this page

Many things only become clear after we begin, not before.

I did not start this blog to prove how much I know, nor to build a machine that has to publish every day. It is closer to a public workbench: a place for questions worth returning to, decisions I have made, mistakes I have learned from, and occasional thoughts I do not want to lose.

Why write in public

Moving from architecture into algorithms was an important change in my learning path, but it is not the blog’s only subject. What matters to me now is recommender systems, language models, agent engineering, and how technical judgement becomes clearer through experiments.

The path to a conclusion is often more useful than the conclusion alone: how the problem was framed, why a hypothesis seemed reasonable, how an experiment was designed, and what a failure changed. Writing in public makes these hidden decisions more concrete and gives my future self a chance to test them again.

Writing may not produce an answer immediately, but it makes vague experience inspectable and gives change a trace. Years from now, I hope this is not only a list of articles, but a map of how I learned to think.

What I will keep here

  • Recommender systems, ranking models, and user-behaviour modelling
  • Learning notes on language models, RAG, agents, and evaluation
  • Engineering practice, debugging, and tools worth sharing
  • Thoughts on interdisciplinary learning and career choices

Start with the first question that is genuinely worth pursuing.

RELATED WRITING

DISCUSSION

Discussion