<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Gustavo Tavares on Medium]]></title>
        <description><![CDATA[Stories by Gustavo Tavares on Medium]]></description>
        <link>https://medium.com/@gustavo600?source=rss-bef8214be5cb------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*ZCXWUVTw_5NcMPsKJPt4_A.jpeg</url>
            <title>Stories by Gustavo Tavares on Medium</title>
            <link>https://medium.com/@gustavo600?source=rss-bef8214be5cb------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 03:51:30 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@gustavo600/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[My 8-Week Study Plan for Mastering Generative AI]]></title>
            <link>https://medium.com/@gustavo600/my-8-week-study-plan-for-mastering-generative-ai-23b715013939?source=rss-bef8214be5cb------2</link>
            <guid isPermaLink="false">https://medium.com/p/23b715013939</guid>
            <category><![CDATA[agents]]></category>
            <category><![CDATA[rags]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[study]]></category>
            <category><![CDATA[llm]]></category>
            <dc:creator><![CDATA[Gustavo Tavares]]></dc:creator>
            <pubDate>Mon, 07 Jul 2025 13:50:05 GMT</pubDate>
            <atom:updated>2025-07-07T14:37:46.093Z</atom:updated>
            <content:encoded><![CDATA[<h3>An engineer’s roadmap from fundamentals to full-stack LLM apps.</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hEiBPb2wdqJrjEfLSGjecg.png" /></figure><p>As a backend software engineer, I wanted a structured path to dive deep into Generative AI. This isn’t about just using APIs; it’s about understanding the core concepts, knowing the architectures, and being able to build, deploy, and evaluate AI-powered applications. I created this 8-week study plan to do just that, dedicating about two hours each weekday.</p><p>This plan is designed to be hands-on, with a mini-project almost every week to solidify the concepts. Here’s the detailed roadmap, enriched with the best resources I could find.</p><h3>📅 Week 1 — Fundamentals of AI and NLP</h3><p>This week is all about building the foundational intuition. As a backend engineer, you’ll appreciate the focus on architecture and data flow.</p><p><strong>Day 1: Neural Networks and Deep Learning</strong></p><ul><li><strong>Course:</strong> <a href="https://www.coursera.org/specializations/deep-learning">Deep Learning Specialization by Andrew Ng (Coursera)</a>. Focus on Course 1: “Neural Networks and Deep Learning.”</li><li><strong>Topics:</strong> Perceptrons, hidden layers, activation functions, backpropagation.</li><li><strong>Extra Reading:</strong> <a href="https://www.deeplearningbook.org/">Deep Learning by Goodfellow, et al.</a> (Chapters 1–2).</li><li><strong>Alternative Visual Explanation:</strong> <a href="https://www.youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_ZCJB-3pi">3Blue1Brown’s series on Neural Networks</a> is a masterpiece for building intuition.</li></ul><p><strong>Day 2: Embeddings</strong></p><ul><li><strong>Article:</strong> <a href="https://jalammar.github.io/illustrated-word2vec/">The Illustrated Word2vec by Jay Alammar</a>. A fantastic visual explanation.</li><li><strong>Explore:</strong> Word2Vec, <a href="https://nlp.stanford.edu/projects/glove/">GloVe</a>, and <a href="https://fasttext.cc/">FastText</a>.</li><li><strong>Practice:</strong> Use <a href="https://radimrehurek.com/gensim/auto_examples/index.html#documentation">Gensim </a>or <a href="https://spacy.io/usage/vectors-similarity">spaCy</a> to load pre-trained vectors and compute similarity.</li></ul><p><strong>Day 3: Tokenization and NLP Pipelines</strong></p><ul><li><strong>Article:</strong> <a href="https://huggingface.co/learn/nlp-course/chapter2/4">How Tokenizers Work (Hugging Face Blog)</a>.</li><li><strong>Tools:</strong> <a href="https://huggingface.co/docs/tokenizers/index">Hugging Face Tokenizers Library</a> and <a href="https://spacy.io/api/tokenizer">spaCy’s Tokenizer</a>.</li><li><strong>Practice:</strong> Try Byte-Pair Encoding (BPE) and WordPiece using the <a href="https://huggingface.co/learn/nlp-course/chapter6/2?fw=pt">Hugging Face “Build a Tokenizer” Tutorial</a>.</li></ul><p><strong>Day 4: Transformers and Self-Attention (Intro)</strong></p><ul><li><strong>Visual:</strong> <a href="http://jalammar.github.io/illustrated-transformer/">The Illustrated Transformer by Jay Alammar</a>. This is the essential starting point.</li><li><strong>Watch:</strong> <a href="https://www.youtube.com/watch?v=FWFA4DGuzSc">Attention is All You Need | A Deep Dive</a> by Letitia Parcalabescu or <a href="https://www.youtube.com/watch?v=U0s0f995w14">StatQuest: The Transformer, illustrated</a> for a simpler breakdown.</li></ul><p><strong>Day 5: Review + Mini Project</strong></p><ul><li><strong>Project:</strong> Build a sentiment analysis pipeline using scikit-learn or Hugging Face.</li><li><strong>Code Example:</strong> <a href="https://huggingface.co/learn/nlp-course/chapter3/3?fw=pt">Hugging Face Tutorial: Fine-tuning a model for sequence classification</a>.</li><li><strong>Goal:</strong> Share a working version on GitHub with a clear README.md.</li></ul><h3>📅 Week 2 — LLMs and Architectures</h3><p>Now we move from general concepts to the specific architectures that power modern Generative AI.</p><p><strong>Day 1: Transformer Architecture Deep Dive</strong></p><ul><li><strong>Read:</strong> <a href="https://arxiv.org/abs/1706.03762"><em>Attention Is All You Need</em> (Original Paper)</a>. Focus on the overall architecture diagram.</li><li><strong>Practice:</strong> <a href="https://www.youtube.com/watch?v=kCc8FmEb1nY">Let’s build GPT: from scratch, in code, spelled out.</a> by Andrej Karpathy. This is a must-watch for any developer.</li></ul><p><strong>Day 2: Modern LLMs (GPT, T5, LLaMA)</strong></p><ul><li><strong>Compare Architectures:</strong> Understand the differences between decoder-only (GPT, LLaMA) and encoder-decoder (T5) models.</li><li><strong>Reading:</strong> <a href="https://ai.meta.com/blog/meta-llama-3/">LLaMA 3 Blog Post</a>, <a href="https://arxiv.org/abs/1910.10683">T5 Paper</a>.</li></ul><p><strong>Day 3: Fine-Tuning vs. LoRA</strong></p><ul><li><strong>Read:</strong> <a href="https://huggingface.co/blog/peft">Hugging Face PEFT Blog</a> and the <a href="https://arxiv.org/abs/2106.09685">LoRA Paper</a>. PEFT (Parameter-Efficient Fine-Tuning) is a critical concept for affordable customization.</li><li><strong>Practice:</strong> Follow the <a href="https://huggingface.co/blog/llama3#fine-tuning-with-peft-and-bitsandbytes">Fine-tuning a Llama 3 model with LoRA</a> tutorial.</li></ul><p><strong>Day 4: Hugging Face Hands-On</strong></p><ul><li><strong>Task:</strong> Use the <a href="https://huggingface.co/docs/transformers/main/en/main_classes/pipelines">Hugging Face Pipelines</a> for zero-shot tasks like question-answering, text-generation, and summarization.</li></ul><p><strong>Day 5: Mini Chatbot Project</strong></p><ul><li><strong>Build:</strong> Create a FastAPI wrapper for a conversational model from Hugging Face.</li><li><strong>Code Example:</strong> <a href="https://prosperasoft.com/blog/artificial-intelligence/rag/fastapi-huggingface-chatbot-integration/">Serving Transformers with FastAPI</a>.</li></ul><h3>📅 Week 3 — RAG (Retrieval-Augmented Generation) and Vector Databases</h3><p>This is where AI meets traditional data retrieval — a concept very familiar to backend engineers.</p><p><strong>Day 1: Semantic Embeddings</strong></p><ul><li><strong>Use:</strong> <a href="https://www.sbert.net/">Sentence-Transformers Library</a> for creating document embeddings.</li><li><strong>Compare:</strong> Explore API-based options like <a href="https://platform.openai.com/docs/guides/embeddings">OpenAI</a> and open-source models like <a href="https://huggingface.co/BAAI/bge-m3">BGE M3</a>.</li></ul><p><strong>Day 2: Vector Databases</strong></p><ul><li><strong>Explore:</strong> Start with a developer-friendly local option like <a href="https://www.trychroma.com/">Chroma DB</a> or a powerful, scalable one like <a href="https://weaviate.io/">Weaviate</a>.</li><li><strong>Read:</strong> <a href="https://www.pinecone.io/learn/vector-database/">What is a Vector Database?</a> from Pinecone’s blog.</li></ul><p><strong>Day 3: RAG Architecture</strong></p><ul><li><strong>Read:</strong> Get familiar with the core frameworks: <a href="https://www.langchain.com/">LangChain</a> and <a href="https://www.llamaindex.ai/">LlamaIndex</a>.</li><li><strong>Paper:</strong> Skim the original <a href="https://arxiv.org/abs/2005.11401">RAG paper</a>.</li></ul><p><strong>Day 4: RAG Pipeline Hands-On</strong></p><ul><li><strong>Build:</strong> Create a full RAG pipeline: Load docs -&gt; Split text -&gt; Embed chunks -&gt; Store in Chroma -&gt; Retrieve and Generate.</li><li><strong>Tutorial:</strong> Follow the <a href="https://python.langchain.com/docs/tutorials/rag/">LangChain RAG Tutorial</a>.</li></ul><p><strong>Day 5: RAG Mini Project</strong></p><ul><li><strong>Example:</strong> Build a dynamic FAQ system using your own set of documents, LlamaIndex, and a local LLM.</li><li><strong>Code Reference:</strong> <a href="https://docs.llamaindex.ai/en/stable/getting_started/starter_example.html">Build a RAG system with LlamaIndex</a>.</li></ul><h3>📅 Week 4 — AI Agents and Tools</h3><p>Agents are the next level of abstraction, where the LLM acts as a reasoning engine to decide which actions to take. This week is about giving the model agency.</p><p><strong>Day 1: Agent Theory (ReAct, Toolformer)</strong></p><ul><li><strong>Read:</strong> The abstract and introduction of the <a href="https://arxiv.org/abs/2210.03629">ReAct paper (Synergizing Reasoning and Acting in Language Models)</a>. This is the foundational concept for many modern agents.</li><li><strong>Watch:</strong><a href="https://www.youtube.com/watch?v=Eug2clsLtFs"> The ReAct Agent Explained</a> for a clear visual breakdown of the “Thought, Action, Observation” loop.</li><li><strong>Explore:</strong> Skim the <a href="https://arxiv.org/abs/2302.04761">Toolformer paper</a> to understand how models can learn to call APIs.</li></ul><p><strong>Day 2: LangChain Agents</strong></p><ul><li><strong>Read:</strong> Read the<a href="https://python.langchain.com/docs/concepts/#agents"> <strong>official LangChain “Concepts” guide</strong></a> to understand the core components: <strong>Tools</strong>, the <strong>Agent</strong>, and the <strong>Agent Executor</strong>.</li><li><strong>Practice:</strong> Build your first agent using one of LangChain’s built-in tools.</li><li><strong>Code Example:</strong> Follow the <a href="https://python.langchain.com/docs/tutorials/agents/"><strong>official LangChain agent tutorial</strong></a> to build an agent that can use tools like TavilySearch and the python_repl tool.</li></ul><p><strong>Day 3: AutoGPT / AutoGen</strong></p><ul><li><strong>Explore:</strong> Understand the concept of autonomous, conversational agents.</li><li><strong>Tool:</strong> Clone and run Microsoft’s <a href="https://github.com/microsoft/autogen">AutoGen</a> locally. It’s a powerful framework for building applications with multiple collaborating agents.</li><li><strong>Analyze:</strong> Run one of their example scripts, like autogen/samples/tool_use/weather.py, and focus on the prompts and the interaction between the UserProxyAgent and AssistantAgent.</li></ul><p><strong>Day 4: Practical Agent Project</strong></p><ul><li><strong>Project:</strong> Build a custom tool for a LangChain agent. Your backend experience is a major asset here.</li><li><strong>Task:<br>1. </strong>Create a simple API using FastAPI that has one or two endpoints (e.g., GET /weather?city=... or GET /movies?genre=...).<br>2. Follow the official LangChain documentation on <a href="https://python.langchain.com/docs/how_to/custom_tools/">How to Create Custom Tools</a> to Wrap Your API as a Tool.<br>3. Create an agent that can intelligently call your custom tool based on a user’s prompt.</li></ul><p><strong>Day 5: Review + Polish Agent</strong></p><ul><li><strong>Refine:</strong> Improve the docstrings and descriptions for your custom tools. The quality of these descriptions directly impacts how well the agent knows when to use them.</li><li><strong>Error Handling:</strong> Add robust error handling to your API and your agent’s logic. What happens if the API call fails or returns no data?</li><li><strong>GitHub:</strong> Push your complete agent project, including the FastAPI tool and the agent itself, to your GitHub.</li></ul><h3>📅 Week 5 — Prompt Engineering and Evaluation</h3><p>Your backend skills in crafting precise instructions and validating outputs are directly transferable here. This week is about mastering the user-facing part of the LLM.</p><p><strong>Day 1: Prompting Strategies</strong></p><ul><li><strong>Read:</strong> <a href="https://platform.openai.com/docs/guides/prompt-engineering/prompt-engineering">OpenAI’s Prompt Engineering Guide</a>. This is an essential overview of the most effective techniques.</li><li><strong>Practice:</strong> Use the OpenAI Playground, Hugging Face, or a local model to experiment with:</li><li><strong>Zero-Shot Prompting:</strong> Asking a question directly.</li><li><strong>Few-Shot Prompting:</strong> Providing 1–3 examples in the prompt.</li><li><strong>Chain-of-Thought (CoT):</strong> Asking the model to “think step by step.”</li><li><strong>Tool:</strong> Explore <a href="https://promptlayer.com/">PromptLayer</a> to see how teams track, manage, and version their prompts.</li></ul><p><strong>Day 2: Advanced Prompting</strong></p><ul><li><strong>Read:</strong> The <a href="https://python.langchain.com/docs/concepts/prompt_templates/"><strong>LangChain documentation on Prompt Templates</strong></a>.</li><li><strong>Practice:</strong> Create dynamic prompt templates that insert variables into your prompts.</li><li><strong>Explore:</strong> Learn about <a href="https://platform.openai.com/docs/guides/function-calling">OpenAI’s Function Calling</a>. This is a more structured and reliable way to get JSON output from a model, which is perfect for API interactions.</li></ul><p><strong>Day 3: LLM Evaluation</strong></p><ul><li><strong>Concepts:</strong> Understand that evaluating LLMs is difficult. For RAG, key metrics include:</li><li><strong>Faithfulness:</strong> Does the answer come from the provided context?</li><li><strong>Answer Relevancy:</strong> Is the answer relevant to the user’s question?</li><li><strong>Context Precision/Recall:</strong> Did you retrieve the right context chunks?</li><li><strong>Tools:<br>- </strong><a href="https://github.com/explodinggradients/ragas">Ragas</a>: The go-to open-source library for evaluating RAG pipelines. Work through their “Getting Started” guide.<br>- <a href="https://github.com/openai/evals">OpenAI Evals</a>: A framework for creating and running evaluations to measure model quality.</li></ul><p><strong>Day 4: Safety &amp; Red Teaming</strong></p><ul><li><strong>Read:</strong> The <a href="https://genai.owasp.org/llm-top-10/">OWASP Top 10 risks for Large Language Model Applications</a>, focusing on “Prompt Injection.”</li><li><strong>Concepts:</strong> Learn the difference between prompt injection and jailbreaking.</li><li><strong>Explore:</strong> Read the system card or safety report for a major model like <a href="https://storage.googleapis.com/deepmind-media/gemini/gemini_1_report.pdf">Google’s Gemini</a> to understand the safety testing they perform.</li></ul><p><strong>Day 5: Prompt Engineering Project</strong></p><ul><li><strong>Project:</strong> Build a reusable prompt toolkit.</li><li><strong>Task:</strong> Create a Python file or module that contains several well-engineered PromptTemplate objects for various tasks like summarization, JSON extraction, grammar correction, and style transfer.</li><li><strong>Goal:</strong> Make this a tool you can import into future projects to standardize your prompts.</li></ul><h3>📅 Week 6 — Deployment and MLOps</h3><p>Leverage your backend and DevOps experience. This week should feel very natural as you take your models from a notebook to production.</p><p><strong>Day 1: APIs with FastAPI</strong></p><ul><li><strong>Task:</strong> Go beyond a basic API. Implement <a href="https://fastapi.tiangolo.com/advanced/custom-response/#streamingresponse">streaming responses in FastAPI</a> to send back tokens as they are generated, providing a better UX.</li><li><strong>Practice:</strong> Integrate a structured logging library like structlog into your FastAPI app to track requests, latency, and model responses.</li></ul><p><strong>Day 2: Local Model Deployment</strong></p><ul><li><strong>Tools:<br>- </strong><a href="https://ollama.com/">Ollama</a>: Install and use Ollama. It’s the easiest way to download and run open-source models like Llama 3 or Mistral on your local machine.<br>- <a href="https://github.com/vllm-project/vllm">vLLM</a>: For higher performance, explore vLLM. Understand its key optimizations like PagedAttention.</li><li><strong>Compare:</strong> Note the trade-offs: Ollama is for ease of use and local development, while vLLM is for high-throughput, production-level serving.</li></ul><p><strong>Day 3: Docker and DevOps</strong></p><ul><li><strong>Task:</strong> Containerize your entire pipeline.</li><li><strong>Project:<br>1. </strong>Create a Dockerfile for your FastAPI application.<br>2. Create a docker-compose.yml file to spin up your FastAPI service <em>and</em> a vector database (like Weaviate or Chroma) together.<br>3. Manage secrets like API keys using environment variables.</li></ul><p><strong>Day 4: Observability and Monitoring</strong></p><ul><li><strong>Tools:</strong> Explore modern LLM observability platforms.</li><li><a href="https://langfuse.com/">Langfuse</a>: Excellent for tracing complex, multi-step chains and agent interactions.</li><li><a href="https://www.helicone.ai/">Helicone</a>: A powerful proxy to monitor costs, latency, and usage for API-based models.</li><li><strong>Practice:</strong> Integrate Langfuse’s SDK into your FastAPI RAG app. Make one API call and watch the entire trace (retrieval, generation) appear on your Langfuse dashboard.</li></ul><p><strong>Day 5: Project: LLM API + Dashboard</strong></p><ul><li><strong>Project:</strong> Combine the week’s work.</li><li><strong>Goal:</strong> Have a fully containerized, observable LLM API. You should be able to run docker-compose up and have a working endpoint that you can monitor in a tool like Langfuse.</li></ul><h3>📅 Week 7 — Real-World Projects and Use Cases</h3><p>This week is about applying your skills to build more complex and useful applications.</p><p><strong>Day 1: Memory-Enabled Chatbot</strong></p><ul><li><strong>Guide:</strong> Read the official <a href="https://python.langchain.com/docs/how_to/chatbots_memory/"><strong>LangChain documentation on how to add memory to chatbots</strong></a>.</li><li><strong>Task:</strong> Build a chatbot that remembers previous turns in the conversation.</li><li><strong>Practice:</strong> Use ConversationBufferMemory to store the history and see how it gets added to the prompt automatically.</li><li><strong>Example:</strong> Build an HR assistant bot that can answer follow-up questions about company policies.</li></ul><p><strong>Day 2: Developer Copilot</strong></p><ul><li><strong>Models:</strong> Explore code-focused models like <a href="https://huggingface.co/codellama">CodeLlama</a> or <a href="https://huggingface.co/bigcode/starcoder2-15b">StarCoder2</a>.</li><li><strong>Practice:</strong> Use one of these models via Hugging Face transformers to generate code, create documentation, and refactor a piece of Python code.</li></ul><p><strong>Day 3: RAG App with UI</strong></p><ul><li><strong>Frontend:</strong> Build a simple, user-friendly interface for one of your previous projects.</li><li><strong>Tools:<br>- </strong><a href="https://streamlit.io/">Streamlit</a>: Incredibly fast for turning Python scripts into beautiful web apps.<br>- <a href="https://www.gradio.app/">Gradio</a>: Also very easy and great for creating demos for any machine learning model.</li><li><strong>Project:</strong> Create a Streamlit app with a text input box and a submit button that displays the answer from your RAG pipeline.</li></ul><p><strong>Day 4: Custom Open Project</strong></p><ul><li><strong>Brainstorm:</strong> Choose a domain you are passionate about (finance, education, gaming, etc.).</li><li><strong>Define:</strong> What problem can you solve with a language model?</li><li>A legal document summarizer?</li><li>A study helper that creates quizzes from textbooks?</li><li>A financial news sentiment analyzer?</li><li>A card game judge that helps you with game rules?</li><li><strong>Plan:</strong> Outline the architecture of your application. What data will you use? What will the user interface look like?</li></ul><p><strong>Day 5: Polish and Publish to GitHub</strong></p><ul><li><strong>Task:</strong> Take one of the projects from this week (the chatbot or the UI app) and polish it for your portfolio.</li><li><strong>Goal:</strong> Write a high-quality README.md with a description, setup instructions, and a GIF or screenshot of the app in action.</li></ul><h3>📅 Week 8 — Final Project and Exploration</h3><p>The final week is for consolidating your knowledge into a single capstone project and looking ahead at more advanced topics.</p><p><strong>Day 1–3: Final Project</strong></p><ul><li><strong>Build:</strong> Execute on the plan from Day 4 of Week 7.</li><li><strong>Goal:</strong> Build a full, “vertical-slice” application that includes:<br>- <strong>Backend:</strong> A FastAPI server with a RAG pipeline and/or an Agent.<br>- <strong>Frontend:</strong> A Streamlit or Gradio UI for user interaction.<br>- <strong>Ops:</strong> The entire application is containerized with Docker.<br>- <strong>Observability:</strong> Integrated logging and tracing with a tool like Langfuse.<br>- <strong>Publish:</strong> Host the project on GitHub with a comprehensive README.</li></ul><p><strong>Day 4: Multimodal Models</strong></p><ul><li><strong>Explore:</strong> Step beyond text-only models.</li><li><strong>Tools:<br>- Vision:</strong> Experiment with a model like <a href="https://llava-vl.github.io/">LLaVA</a> for asking questions about images.<br>- <strong>Audio:</strong> Use the <a href="https://platform.openai.com/docs/guides/speech-to-text">OpenAI Whisper API</a> to transcribe audio into text.<br>- <strong>Image Generation:</strong> Use the <a href="https://platform.openai.com/docs/guides/images">DALL·E 3 API</a> to generate images from text prompts.</li><li><strong>Practice:</strong> Build a simple script that takes a voice command via Whisper and uses it to generate an image with DALL·E 3.</li></ul><p><strong>Day 5: Reflection + Next Steps</strong></p><ul><li><strong>Review:</strong> Look back at your GitHub. See how far you’ve come from the simple scripts of Week 1 to your full-stack app in Week 8.</li><li><strong>Write:</strong> Publish a blog post on Medium or a detailed project summary on LinkedIn. Share your learnings, challenges, and insights.</li><li><strong>Plan:</strong> What’s next?</li><li>Contribute to an open-source project like LangChain, LlamaIndex, or Ollama.</li><li>Deepen your knowledge in a specific area (like evaluation or fine-tuning).</li></ul><p>This structured approach should provide a solid foundation for anyone with a software engineering background looking to become proficient in building and deploying real-world Generative AI applications. Good luck!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=23b715013939" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>