<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>RenoCrypt · Matrix Alchemy</title><description>Writing on machine learning, AI, and engineering from RenoCrypt.</description><link>https://www.renocrypt.com/</link><language>en</language><item><title>Math Foundations of Transformers and MoE Layers</title><link>https://www.renocrypt.com/dev/math-behind-transformers-moe/</link><guid isPermaLink="true">https://www.renocrypt.com/dev/math-behind-transformers-moe/</guid><description>A thorough explanation of the equations powering classic transformer structures and Mixture-of-Experts for advanced deep learning workflows.</description><pubDate>Sun, 13 Apr 2025 00:00:00 GMT</pubDate><category>ai systems engineering</category><category>transformers</category><category>mixture of experts</category><category>machine learning</category><category>deep learning</category><category>neural networks</category><category>model architecture</category><category>computational math</category></item><item><title>Self-Hosting OpenWebUI: A Step-by-Step Walkthrough</title><link>https://www.renocrypt.com/dev/openwebui-server-self-host/</link><guid isPermaLink="true">https://www.renocrypt.com/dev/openwebui-server-self-host/</guid><description>A technical guide for deploying a production-ready OpenWebUI instance on your own domain with Docker, Nginx, and Let&apos;s Encrypt for secure AI interactions.</description><pubDate>Sat, 12 Apr 2025 00:00:00 GMT</pubDate><category>ai systems engineering</category><category>openwebui</category><category>self hosting</category><category>docker</category><category>nginx</category><category>lets encrypt</category><category>ai interface</category><category>devops</category><category>infrastructure</category><category>security</category><category>containerization</category></item><item><title>Production-Ready AI Agents: Architectures, Frameworks, and Beyond</title><link>https://www.renocrypt.com/dev/agent-ai/</link><guid isPermaLink="true">https://www.renocrypt.com/dev/agent-ai/</guid><description>Production-grade AI agent systems: the frameworks, orchestration patterns, and deployment strategies behind autonomous systems that hold up at scale.</description><pubDate>Tue, 04 Feb 2025 00:00:00 GMT</pubDate><category>ai systems engineering</category><category>ai agents</category><category>autonomous systems</category><category>langchain</category><category>autogen</category><category>crewai</category><category>rag</category><category>orchestration</category><category>production ml</category><category>function calling</category><category>system design</category></item><item><title>Enhancing LLM Reasoning: Frameworks and Methodologies</title><link>https://www.renocrypt.com/dev/llm-reasoning/</link><guid isPermaLink="true">https://www.renocrypt.com/dev/llm-reasoning/</guid><description>Chain-of-Thought, Tree-of-Thoughts, and ReAct compared as ways to give large language models structured reasoning rather than a single guess.</description><pubDate>Mon, 03 Feb 2025 00:00:00 GMT</pubDate><category>machine learning fundamentals</category><category>large language models</category><category>reasoning</category><category>chain of thought</category><category>tree of thoughts</category><category>react</category><category>prompt engineering</category></item><item><title>Understanding Tokenizers and Embedders in LLM Pipelines</title><link>https://www.renocrypt.com/dev/tokenizers-and-embedders/</link><guid isPermaLink="true">https://www.renocrypt.com/dev/tokenizers-and-embedders/</guid><description>A deep dive into the role, structure, and training of tokenizers and embedders in modern language models like GPT, BERT, and T5.</description><pubDate>Sun, 26 Jan 2025 00:00:00 GMT</pubDate><category>machine learning fundamentals</category><category>machine learning</category><category>transformers</category><category>tokenizers</category><category>large language models</category></item><item><title>Trust Region Policy Optimization, Proximal Policy Optimization</title><link>https://www.renocrypt.com/posts/rl-more-policy-gradient/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/rl-more-policy-gradient/</guid><description>How TRPO and PPO keep policy updates stable in reinforcement learning, and where off-policy methods reuse past experience to get there.</description><pubDate>Tue, 20 Aug 2024 00:00:00 GMT</pubDate></item><item><title>GPT, PaLM, GLM, LLaMA and beyond: Large language models in medicine</title><link>https://www.renocrypt.com/gpt-palm-glm-llama-and-beyond-large-language-models-in-medicine/</link><guid isPermaLink="true">https://www.renocrypt.com/gpt-palm-glm-llama-and-beyond-large-language-models-in-medicine/</guid><description>GPT, PaLM, GLM, and LLaMA applied to medicine: where they are already useful, where they fall short, and what the open challenges are.</description><pubDate>Sun, 18 Aug 2024 00:00:00 GMT</pubDate><category>machine learning theory</category><category>large language models</category><category>machine learning</category><category>machine learning models</category></item><item><title>Power of Multimodal Machine Learning: Challenges, Innovations, and Future Directions</title><link>https://www.renocrypt.com/multimodal-machine-learning-challenges-innovations/</link><guid isPermaLink="true">https://www.renocrypt.com/multimodal-machine-learning-challenges-innovations/</guid><description>What makes multimodal machine learning hard, which recent innovations moved it forward, and where integration across modalities is heading.</description><pubDate>Sat, 17 Aug 2024 00:00:00 GMT</pubDate><category>machine learning theory</category><category>multimodal</category><category>machine learning</category><category>machine learning models</category></item><item><title>Understanding Key Insights and Trends in Large Language Models</title><link>https://www.renocrypt.com/understanding-key-insights-and-trends-in-large-language-models/</link><guid isPermaLink="true">https://www.renocrypt.com/understanding-key-insights-and-trends-in-large-language-models/</guid><description>How large language models are built and how they have evolved, traced through the key insights and trends behind each generation.</description><pubDate>Thu, 15 Aug 2024 00:00:00 GMT</pubDate><category>machine learning theory</category><category>large language models</category><category>machine learning</category><category>machine learning models</category></item><item><title>What is Deep Reinforcement Learning (RL)</title><link>https://www.renocrypt.com/posts/reinforcement-learning/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/reinforcement-learning/</guid><description>The language and notation of reinforcement learning, and what an RL agent actually does, introduced from first principles.</description><pubDate>Fri, 09 Aug 2024 00:00:00 GMT</pubDate><category>machine learning</category><category>reinforcement learning</category><category>theory</category></item><item><title>Gemini 2.0 Pro vs. Flash: A Deep Technical Comparison for Enterprise AI</title><link>https://www.renocrypt.com/dev/gemini-2-pro-flash-technical-comparison/</link><guid isPermaLink="true">https://www.renocrypt.com/dev/gemini-2-pro-flash-technical-comparison/</guid><description>Gemini 2.0 Pro against Flash on architecture, benchmarks, and deployment cost, with a clear rule for choosing between them in enterprise work.</description><pubDate>Sat, 27 Jul 2024 00:00:00 GMT</pubDate><category>ai model engineering</category><category>gemini</category><category>gemini pro</category><category>gemini flash</category><category>large language models</category><category>multimodal</category><category>google ai</category><category>model architecture</category><category>enterprise ai</category><category>inference optimization</category><category>ai benchmarks</category><category>context window</category><category>production ml</category><category>performance</category><category>efficiency</category></item><item><title>An in-Depth Look at Leading Computer Vision ML Models</title><link>https://www.renocrypt.com/an-in-depth-look-at-leading-computer-vision-models/</link><guid isPermaLink="true">https://www.renocrypt.com/an-in-depth-look-at-leading-computer-vision-models/</guid><description>The leading computer vision model families in one place, from the architecture each uses to the task it is actually good at.</description><pubDate>Sat, 20 Jul 2024 00:00:00 GMT</pubDate><category>machine learning theory</category><category>computer vision</category><category>machine learning</category><category>machine learning models</category></item><item><title>What is OpenMMLab and how it is used in computer vision</title><link>https://www.renocrypt.com/posts/what-is-openmm/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/what-is-openmm/</guid><description>OpenMMLab as a computer vision platform: what MMDetection, MMSegmentation, MMPose, MMOCR, and MMDeploy each cover, and where they fit in a workflow.</description><pubDate>Mon, 20 May 2024 00:00:00 GMT</pubDate><category>machine learning theory</category><category>object detection</category><category>machine learning</category></item><item><title>YOLO and object detection</title><link>https://www.renocrypt.com/various-yolo-versions-and-differences/</link><guid isPermaLink="true">https://www.renocrypt.com/various-yolo-versions-and-differences/</guid><description>How the YOLO object detection series changed from v1 onward, and what each version actually improved over the one before it.</description><pubDate>Wed, 15 May 2024 00:00:00 GMT</pubDate><category>machine learning</category><category>computer vision</category></item><item><title>Yet another gentle intro to the YOLO-NAS model</title><link>https://www.renocrypt.com/posts/yolo-nas/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/yolo-nas/</guid><description>How Deci AI&apos;s YOLO-NAS combines neural architecture search with YOLO object detection, from its multi-phase training to reading the model&apos;s output.</description><pubDate>Mon, 29 Apr 2024 00:00:00 GMT</pubDate><category>machine learning theory</category></item><item><title>Fooocus Fantasia: Navigating Generative Magic</title><link>https://www.renocrypt.com/posts/fooocus-and-how/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/fooocus-and-how/</guid><description>A working guide to Fooocus, the image generation interface that hides the tuning, and how to run it locally or on Colab.</description><pubDate>Tue, 19 Sep 2023 00:00:00 GMT</pubDate><category>machine learning application</category></item><item><title>Behind Generative Models [In Progress]</title><link>https://www.renocrypt.com/posts/generative-models/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/generative-models/</guid><description>The mechanics behind diffusion models, and what separates DALL-E, Midjourney, and Stable Diffusion once you look past the interface.</description><pubDate>Mon, 18 Sep 2023 00:00:00 GMT</pubDate></item><item><title>The Generative Lens: Crafting Camera Angles with Precision</title><link>https://www.renocrypt.com/posts/img-angle-advantage/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/img-angle-advantage/</guid><description>A guide to refining camera angle outputs in generative platforms using nuanced prompt strategies.</description><pubDate>Mon, 11 Sep 2023 00:00:00 GMT</pubDate></item><item><title>Q-learning</title><link>https://www.renocrypt.com/posts/rl-q-learning/</link><guid isPermaLink="true">https://www.renocrypt.com/posts/rl-q-learning/</guid><description>Q-learning is a reinforcement learning technique where an agent iteratively learns the value of its actions to navigate towards optimal outcomes in an environment.</description><pubDate>Sun, 10 Sep 2023 00:00:00 GMT</pubDate><category>machine learning</category><category>reinforcement learning</category><category>theory</category></item></channel></rss>