I am following the lead of other members and posting my progress. Hopefully it will help keep me accountable. I've just completed the Alchemy Codex course! Even with prior experience, it's always great to gain new insights and add fresh tweaks and features to my setup and workflow. Can't wait to jump into the next course!

I wanted to apply what I'm learning in the "Building Applications with LLMs" course without worrying too much about costs, so I started running models locally using Ollama. Since I'm just prototyping, I dont need top-tier models, Ill save those for production (one day). My setup isn't the fastest, so things were a bit slow at first, but learning about quantization really helped me improve performance. Hopefully this helps others in a similar situation. This video explained it well for me for starters, maybe for you too.

A few days ago, I shared that I was working on my first AI agent for a transport and logistics company. Today, that agent is fully operational — managing real-world tasks like spare parts tracking, work order generation, and automated reporting, all via WhatsApp. This project allowed me to integrate GPT with Supabase databases, automation workflows, PDF generation, and external APIs, resulting in a robust solution that’s easy to use even for non-technical teams. If you work with a business (or client) that could benefit from a similar system — or if you need a custom GPT agent tailored to your specific workflows — I’m open to collaborating, regardless of your location. 📩 Feel free to reach out via direct message here on Skool or email me at juaninovara@gmail.com I’m focused on helping businesses scale and modernize through real, functional AI solutions.

I've spent the last year watching something interesting happen in our field... Everyone's talking about GenAI. Every tweet, every product launch, every investor memo. But here's what's really happening: When you zoom in on the AI systems actually running in production today, almost none of them are GenAI. They're still powered by traditional ML — logistic regression, XGBoost, neural nets. But that won't last. I predict those numbers will completely flip over the next decade. GenAI is going to eat everything. But honestly, most developers aren't ready for that shift. Not because they aren't good - but because GenAI is evolving faster than any tech wave we've ever seen. There's no real roadmap. No clean, proven path to go from "I can prompt ChatGPT" to "I build production-ready GenAI systems." I've felt this gap myself. Even with a strong background in AI, I had to dig for answers, hack things together, and experiment like crazy to keep up. And when I started hiring engineers at Datalumina, I realized there's no proper training program for this new wave of AI engineering. So I built one. It started as an internal roadmap. Then I shared it within our community. The response? "This is exactly what I needed!" So now, it's official: The GenAI Accelerator is open for enrollment. ✓ A 6-week cohort-based program ✓ Built for developers who want to level up fast ✓ Focused on real, production-grade systems — not playground projects ✓ Based on 10+ years of building AI systems ✓ First cohort starts May 5th You'll learn techniques that aren't shared on YouTube or Medium. These are the same approaches used by big tech companies, but their employees can't openly share them online due to confidentiality agreements. But I don't care about competition. The AI market is growing exponentially. There's more than enough opportunity for every skilled developer. At Datalumina, we can't even take on 1% of the project requests coming our way. The demand for AI engineers is so massive that I'm not worried about "giving away secrets" - I'm focused on helping more developers build the skills needed to meet this demand.

Kaggle Introduction to ML done! Interesting, inspiring, I want more! 🤩