Learn Fundamentals Before Going Deep: A Gentle Reminder

🧠 Setting Up LangChain 1.0 New Version of LangChain If you're starting with LangChain 1.0, here’s the clean, no-fluff setup guide — exactly what you need, nothing extra 👇 🚀 1️⃣ Prerequisites Use Python 3.11–3.13 ✅ Install uv (superfast package manager) or pip Install Node.js (brew install node or use NodeSource for Ubuntu) 🔑 2️⃣ Get Your API Keys OpenAI API Key → required LangSmith API Key → optional (for tracing & Studio) Add them to a .env file: OPENAI_API_KEY=your_key LANGSMITH_API_KEY=your_key LANGSMITH_TRACING=true 📦 3️⃣ Install LangChain Essentials Repo git clone https://lnkd.in/d8uRextw cd lca-langchainV1-essentials/python cp example.env .env uv sync # or pip install -r requirements.txt 💻 4️⃣ Run Jupyter Labs uv run jupyter lab Optional: langgraph dev # for LangSmith Studio 🧩 5️⃣ Learn by Doing Start with: L1_fast_agent.ipynb → Build your first SQL Agent Then explore: Memory, Tools, Streaming, HITL, Dynamic prompts ✨ Tip: Use uv for dependency management — it’s 10× faster than pip and perfectly fits LangChain workflows. #LangChain #LLM #AIEngineering #Python #MLOps #LangSmith #uv #LLMOps #Developers

HTML SERIES: Episode 8 – Link Me Up, Web! 🔗 Let’s talk about the internet’s favorite magic trick: hyperlinks. Links are the lifeblood of the web because they let us jump between pages, stalk our crush’s blog, or binge late-night tutorials. In HTML, we use the <a> tag (short for anchor) to create them. Basic syntax: <a href="https://example.com">Click me</a> - href = the destination (aka where the link takes you) - The text between <a> and </a> = what people will click Example: <a href="https://techtots.africa">Visit TechTots</a> Open in a new tab? Add: target=_blank <a href=https://techtots.africa target=_blank>Visit TechTots</a> Now you’re not just building a site you’re connecting the entire web. Next episode: Lists because chaos is fun, but order is cleaner. #HTMLLinks #LearnToCode #TechTots #FrontendDev #HyperlinkEverything #DigitalCreatives #LinkedInLearning #HTMLSeries

🚀 𝗠𝗮𝗿𝗸𝗱𝗼𝘄𝗻 𝗚𝘂𝗶𝗱𝗲 𝗥𝗲𝗽𝗼𝘀𝗶𝘁𝗼𝗿𝘆! 📖 Whether you're writing a GitHub README, documenting a project, blogging, or taking notes—Markdown is an essential skill every developer, student, and tech enthusiast should know. So, I created a complete Markdown Guide to help beginners learn and practice Markdown in a structured and hands-on way. ✅ 💡 𝗪𝗵𝗮𝘁’𝘀 𝗶𝗻𝘀𝗶𝗱𝗲 𝘁𝗵𝗲 𝗿𝗲𝗽𝗼? ✅ Full Markdown syntax explained ✅ Step-by-step examples ✅ Real-world usage cases ✅ Quick Cheatsheet for daily use 📚 𝗧𝗼𝗽𝗶𝗰𝘀 𝗖𝗼𝘃𝗲𝗿𝗲𝗱 🔹 Headings & Paragraphs 🔹 Bold / Italic / Strikethrough 🔹 Lists (Ordered & Bullet) 🔹 Links, Images & Emojis 🎉 🔹 Code (Inline & Block) 🔹 Tables & Task lists 🔹 Blockquotes & more! 🛠️ 𝗚𝗲𝘁 𝘀𝘁𝗮𝗿𝘁𝗲𝗱 𝗶𝗻 𝗺𝗶𝗻𝘂𝘁𝗲𝘀: git clone https://lnkd.in/djf_kvBE cd Markdown-Guide 🙌 𝗖𝗼𝗻𝘁𝗿𝗶𝗯𝘂𝘁𝗶𝗼𝗻𝘀 𝗪𝗲𝗹𝗰𝗼𝗺𝗲! Have cool examples or improvements? Feel free to submit a PR—I’d love to collaborate! 🔗 𝗥𝗲𝗽𝗼 𝗟𝗶𝗻𝗸: (https://lnkd.in/dDsBFXJn) If you're getting started with documentation or want to improve your GitHub READMEs, this guide is for you. 💙 Would love your feedback and support! #markdown #opensource #github #developers #learning #documentation #readme

Lost trying to find MS Word documents? Where are they? What do they contain? Enter the plain text, structured runbook. Blog Post: https://lnkd.in/eV_ypqzq Financial Services is still using antiquated methods of writing business critical documentation. Recently, I’ve become a fan of the runbook - a short, concise set of instructions to perform a given task. Once you start using these plain text based runbooks going back to a Word document simply feels slow and cumbersome. 👍 Why is it better? Search documents without opening them using filesystem tools. Version control with review process using Git. Generate HTML documents using tools like MK Docs.

After years of writing technical blogs, I realized something was missing. The commands, the diagrams, the benchmarks — they explain how things work, but not why we make them that way. The truth is, most of what we really learn doesn’t come from perfect platforms. It comes from tuning queries that shouldn’t need tuning, rethinking how data flows across storage and compute, and realizing that no amount of automation fixes a bad design. That’s why I started 𝗧𝗵𝗲 𝗔𝗜 𝗟𝗮𝗸𝗲𝗵𝗼𝘂𝘀𝗲 𝗗𝗼𝗷𝗼, a space to share what doesn’t fit in a tutorial: The real-world lessons, the context behind decisions, and what I wish I’d known earlier. No hype. No buzzwords. Just knowledge that comes from doing the work. 🥋 If that’s the kind of learning you care about, join the dojo 👉 https://lnkd.in/dA5vz_5b

Off the back of a successful AI World Oracle has launched its AI Data Platform and Autonomous AI Lakehouse. Together they combine open formats like Iceberg with unified governance, zero-ETL and zero-copy access, cross-cloud interoperability and embedded AI to help teams move from data to insight faster. My colleague Javier de la Torre Medina has created the AI Lakehouse DOJO. It dives into practical real life examples of how Oracle has build a gen2 Cloud for Enterprises, that's creating true differential to other hyperscalers. Join the DOJO here: https://lnkd.in/e3dRHGn9

What is a repository "Fork"? A fork is your own copy of someone else’s repository. It lets you explore, modify, and experiment without affecting the original project. Why fork? * Safe experimentation - break things freely in your own copy * Contribute easily - make changes and submit a pull request * Customize - adapt open-source projects to your needs * Learn - study real-world codebases hands-on

🚀 FastAPI Learning – Day 1 Today I kicked off a new learning project using FastAPI (Python framework for building APIs). Instead of jumping into complex features, I focused on doing the fundamentals properly: Set up the project with a virtual environment Installed FastAPI & Uvicorn Created a simple route (/) and tested it with Swagger UI Cleaned the repository by removing venv/ and __pycache__ Added a proper .gitignore Pushed the code to GitHub with clear commits 🔧 Tech used: Python FastAPI Uvicorn Git & GitHub for version control 📌 Repo link: 👉 https://lnkd.in/dFNvtTFD On the surface, this is a simple setup, but it matters because: If the foundation is messy, the entire project becomes chaos later. Tomorrow, I’ll start working on API models and database integration (SQL). If anyone here has built production APIs with FastAPI, I’d love to hear your advice on best practices.

🛠️ Built Markify – A Chrome Extension for PDF Annotation Just finished building a Chrome extension that enhances PDF handling in the browser. Here's what's under the hood: Core Features: • In-browser PDF rendering with custom viewer • Real-time annotation and highlighting tools with color picker option • Sticky notes for document markup • Persistent annotations across sessions (saved locally) Tech Stack: • PDF.js for document rendering • pdf-lib.js for PDF manipulation • Vanilla JavaScript for the extension logic • Chrome Extension Manifest V3 • Custom CSS for the viewer interface Architecture: • Background service worker for PDF detection • Content scripts for seamless integration • Custom viewer with modular components • Event-driven communication between scripts Built from scratch without heavy frameworks – focused on performance and lightweight footprint. The extension intercepts PDF requests and renders them in a custom viewer with annotation capabilities baked in. 🔗 GitHub: https://lnkd.in/gasbDewn #WebDevelopment #ChromeExtension #JavaScript #PDFjs #OpenSource #SoftwareEngineering #Frontend

Stuck in tutorial hell? Here’s why: You sit through a 3-hour tutorial. Everything makes sense while you’re watching. Then you close the video… and suddenly, your mind goes blank. The problem isn’t that you can’t code, it’s that you learned to follow instructions, not to build independently. Here’s how I fixed it👇 1. Watch 10 minutes of a tutorial. 2. Pause it. Close it. 3. Try to rebuild what you just learned from memory. 4. When you get stuck, Google the specific error or question. 5. Make it work your way. Do this 30 times, and you’ll have 30 mini-projects you actually understand, not just ones you copied. That’s how you really learn to code. Stop watching endlessly. Start building intentionally. 💻