Today, I had the opportunity to attend the “Master Agentic AI” live masterclass hosted by Mohammad Arshad from Decoding Data Science. What stood out to me was how the session moved beyond the hype around LLMs and focused on what it actually takes to build production-grade AI agents. 💡 Key takeaways: 🔹 Architecture matters. We explored the architecture of HALA, an HR AI agent designed to handle annual leave allocation for corporates. It was a great example of how agentic AI can address a real-world business workflow rather than simply generate responses. 🔹 CPE — Clarity, Prototype, Evidence. Mr. Arshad introduced the CPE framework, highlighting the importance of starting with clarity around the problem and objective, building a prototype, and then gathering evidence to validate whether the solution actually works. 🔹 Clarity can matter more than the model. One of the most valuable lessons for me was that the success of a production AI system isn't necessarily determined by choosing the “best” LLM. Clear problem definition, architecture, data strategy, and implementation decisions can have a much greater impact. 🔹 Chunking strategy matters. Something as seemingly technical as choosing the right chunking strategy for data can significantly influence the quality and reliability of a production-grade AI system. The lesson was clear: better architecture and better data preparation can matter more than simply switching to a more powerful model. 🔹 Build for evidence, not just a demo. A prototype may look impressive, but production readiness requires validation, evaluation, reliability, and evidence that the system performs well against the actual business requirements. 🏆 And a personal highlight: I’m happy to share that I secured 1st place in the Kahoot quiz during the session! 🎉 A big thank you to Mohammad Arshad and Decoding Data Science for such an insightful and practical masterclass. My biggest takeaway: Don’t start with “Which LLM should I use?” Start with “What problem am I solving, how should I architect the solution, and what evidence will prove that it works?”

Just took the DDS Builder Codex Pledge. 🚀 For me, this isn’t just about making a promise—it’s about becoming more intentional about the person and builder I want to be. The pledge reminded me that progress comes from identity, learning, building, and community. It’s easy to stay in planning mode, wait until everything feels perfect, or let fear of failure slow you down. My commitment is to keep showing up, keep learning, build consistently, and share the journey with the community. I’m choosing action over hesitation, progress over perfection, and consistency over excuses. Excited to see where this commitment takes me—and to build alongside all of you. 🤝 I took the pledge. Now it’s time to live it. 🔥

Receiving this winner’s certificate is a proud milestone in my journey of developing ProSightAI into a fully functional, secure, and responsible Agentic AI application. A huge thank you to Decoding Data Science, the organizers, mentors, judges, and everyone in this community for the guidance, encouragement, and opportunity to learn and build alongside other AI enthusiasts. This achievement motivates me to keep improving ProSightAI, strengthen its agentic and security capabilities, and turn it into an even more impactful solution. Grateful for the journey—and excited for what comes next! 🚀🤖