OpenAI just dropped a major update: the Responses API. Whenever OpenAI releases something like this, it changes the game for developers, forcing us to rethink how we build AI applications. In this week’s video, I break down exactly what this API does, what’s changing, and whether you should migrate your projects. Key updates: - It’s a superset of the Chat Completions API (meaning it does everything Chat Completions did—plus more). - New built-in tools: Web search, file search, and computer use. - Simplified API calls, but also more abstraction—which can be both good and bad. If you’re serious about staying ahead in AI development, you’ll want to watch this. Check out the full breakdown here

A 6 weeks interactive course to learn how to build and deploy your own agents. Register for free 👇🏼https://huggingface.co/agents-course

When building AI agents, you need them to understand your data—whether it’s PDFs, websites, or internal documents. Most tools for this are closed-source, requiring API keys and external platforms. But what if you could do it all in Python with an open-source library? In this week’s video, I show you how to build a fully open-source document extraction pipeline using Docling. You’ll learn how to: - Extract, parse, and chunk documents for AI processing. - Store and retrieve data efficiently with vector databases. - Build a working chat application that can answer questions based on your documents. Watch the video here.

These courses, from DeepLearning AI, are really complementing very well each other and are a GOLDMINE - How transformer LLMs work - Attention in Transformers: Concepts and Code in PyTorch The first one involves Jay Alammar and Maarten Grootendorst, co-authors of the book "Hand-On Large Language Models" (the "framework" followed being HuggingFace) What makes the book and the courses shine are the relevant illustrations. By the way, they're continually adding resources to the book, via their YouTube channels (see above) and their Newsletter (their last article being a "debunking via illustration" of DeepSeek's architecture) The second one involves Joshua Starmer (💥💥💥) from the reknowned StatsQuest channel and LightningAI He has also written several great visual books related to Machine Learning ENJOY!

Hey everyone! As we’re heading towards the final month of 2024, I’ve been reflecting on what a whirlwind the past few months have been. Q3 and Q4 have been incredibly busy — between running Datalumina, managing client projects, seeing amazing growth in Data Freelancer, shipping the GenAI launchpad, moving into a new office, and onboarding the first customers for our new SaaS product, it’s been a ride. 🚀 On top of that, keeping up with the YouTube channel has been both a challenge and a joy. Honestly, it’s one of my favorite parts of what I do — helping you all learn, grow, and tackle data challenges with confidence. Seeing your feedback and progress keeps me motivated to keep creating. Now, I’m starting to map out my content plans for 2025, and I’d love to hear from YOU. What are you currently learning? What topics would you like me to cover next year? 🎥 Would you like to see more LLM-based tutorials? Deeper dives into machine learning workflows? Maybe even some freelancing tips for data professionals? Or something else entirely? Let me know in the comments — what are you most excited to learn in 2025? This is your chance to shape the direction of the content we create together. Thanks for being part of this community! 🙏🏻 — Dave