<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>LLM Development on AI VOID</title><link>https://ai-blog.noorshomelab.dev/tags/llm-development/</link><description>Recent content in LLM Development on AI VOID</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sat, 11 Apr 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://ai-blog.noorshomelab.dev/tags/llm-development/index.xml" rel="self" type="application/rss+xml"/><item><title>Weakly Supervised Distillation of Hallucination Signals into Transformer Representations: Research Explainer for Builders</title><link>https://ai-blog.noorshomelab.dev/research/weakly-supervised-hallucination-distillation/</link><pubDate>Sat, 11 Apr 2026 00:00:00 +0000</pubDate><guid>https://ai-blog.noorshomelab.dev/research/weakly-supervised-hallucination-distillation/</guid><description>&lt;h2 id="quick-verdict"&gt;Quick Verdict&lt;/h2&gt;
&lt;p&gt;Hallucination is the Achilles&amp;rsquo; heel of Large Language Models (LLMs). This paper presents a compelling new approach that moves beyond external fact-checking to make LLMs &lt;em&gt;internally aware&lt;/em&gt; of their own potential hallucinations. By distilling weak, noisy signals into the model&amp;rsquo;s hidden representations during training, it aims to create LLMs that can inherently distinguish between factual and fabricated information at a deeper level. For developers building reliable LLM applications, this is a significant step towards more trustworthy and self-aware AI.&lt;/p&gt;</description></item><item><title>A Comprehensive Guide to Teach me a complete step-by-step career path for Applied AI and Agentic AI development, starting from foundational programming and system thinking, then moving into working with large language models and AI APIs, prompt engineering, tool use, function calling, retrieval-augmented generation (RAG), memory and state management, agent orchestration, multi-agent systems, AI-driven workflows, evaluation and observability, cost and latency optimization, security and privacy considerations, and production deployment, with a strong focus on building real applications that use AI at its full potential, including progressively challenging hands-on projects, daily practice ideas, system design patterns, common failure modes, and sections that encourage independent experimentation and idea generation so I can grow from beginner to professional applied AI engineer and product builder, aligned with modern agentic AI practices as of January 2026. Chapters</title><link>https://ai-blog.noorshomelab.dev/applied-agentic-ai-2026-guide/</link><pubDate>Fri, 16 Jan 2026 00:00:00 +0000</pubDate><guid>https://ai-blog.noorshomelab.dev/applied-agentic-ai-2026-guide/</guid><description>&lt;p&gt;Explore a comprehensive collection of chapters designed to guide you from beginner to professional in Applied AI and Agentic AI development. This path covers everything from foundational programming to advanced agent orchestration and production deployment, with a strong focus on building real-world AI applications. Discover progressively challenging projects, system design patterns, and expert insights to master modern AI practices.&lt;/p&gt;</description></item></channel></rss>