I used to think a personal blog was mostly a publishing problem: pick a static site framework, move some notes over, and write when I had time.

That was too small a frame.

What I need is closer to a public research notebook. I am still early in my path, so this site should not pretend to be an academic CV or a list of finished results. It should show the work in formation: the concepts I keep returning to, the systems I build to think with, and the notes that survive more than one week of curiosity.

My long-term center of gravity is AI research, especially representation learning, explainability, ontological systems and world models, and the mathematics and linguistics underneath learning systems. Those interests do not fit neatly into one folder. A theorem note, a paper reading trail, a small engineering system, and a bilingual reflection can all be part of the same thinking process.

That is why the site has two main surfaces.

Blog posts are for ideas that have enough shape to be argued. They can still be personal, but they should make a claim, tell a learning story, or explain a technical design choice.

Notes are smaller knowledge units. Some are rough. Some are just seeds. The point is that they are durable enough to be linked, revised, translated, and eventually connected in the graph.

The implementation details matter only because they support that workflow. I want stable URLs, bilingual counterparts, mathematical environments, comments, search, and a graph that makes relationships visible. I also want room to add more academic elements later, such as a CV, publication pages, and project pages, without turning the homepage into a formal CV before I have the research output to justify it.

So this site is not a portfolio in the usual software-engineering sense. It is a place where my research identity can become legible over time.