<?xml version="1.0" encoding="utf-8"?><feed xmlns="http://www.w3.org/2005/Atom" ><generator uri="https://jekyllrb.com/" version="3.10.0">Jekyll</generator><link href="https://belatijagad.github.io/feed.xml" rel="self" type="application/atom+xml" /><link href="https://belatijagad.github.io/" rel="alternate" type="text/html" /><updated>2026-07-10T23:06:58+07:00</updated><id>https://belatijagad.github.io/feed.xml</id><title type="html">Belati Jagad</title><subtitle>Entry-level ML engineer and data scientist with research experience in language-model uncertainty, preference optimization, and multilingual representation learning.</subtitle><author><name>Belati Jagad</name><email>belatijagadbintangsyuhada@gmail.com</email></author><entry><title type="html">Bitter Lesson on LLM Alignment Research from an Undergrad</title><link href="https://belatijagad.github.io/notes/alignment-bitter-lesson/" rel="alternate" type="text/html" title="Bitter Lesson on LLM Alignment Research from an Undergrad" /><published>2026-07-01T00:00:00+07:00</published><updated>2026-07-01T00:00:00+07:00</updated><id>https://belatijagad.github.io/notes/alignment-bitter-lesson</id><author><name>Belati Jagad</name><email>belatijagadbintangsyuhada@gmail.com</email></author><category term="notes" /><category term="llm" /><category term="alignment" /><category term="sft" /><category term="dpo" /><category term="rl" /><summary type="html"><![CDATA[A reflection on two undergraduate research internships at MBZUAI and the practical tension between simple alignment ideas, expensive experiments, and uncertain results.]]></summary></entry></feed>