Mohammad Arshad (@arshad83) header
Mohammad Arshad (@arshad83) profile picture

Principal Data Scientist | 120K+ Linkedin Community ! Generative AI | 20 Years+ Exp | Ex-Accenture, HP, Dell | Keynote Speaker & Mentor | LLM, AWS, Azure GCP

AI Agents & RAG Tutorials →
Joined April 2009
Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)Photo by Mohammad Arshad (@arshad83)
Supercharge Your AI Skills for 2025 Unlock $2,000 worth of resources — including free courses, recordings, and ebooks on Generative AI, Agentic AI, and MCP. 👉 Limited time access. To get 1⃣ Follow(for DM access) 2⃣Like and Repost this post 3️⃣Reply with "Send" – at United Arab Emirates
203
172
5
405
31,854
GPT-6.2 is here. What actually changed: - Two models: Astra (frontier) + Sol (everyday work) - Astra: $8 in / $40 out per 1M tokens, 20% cheaper than GPT-6 Astra - Sol: $2 in / $10 out, same price as 6.1 A new bar for professional work. Which one are you building with first?
37
Ten tabs became one AI chat. My real setup: - Summarise 200 emails in seconds - Draft X posts for my approval - Check community numbers - Query my database + Google Forms Built with MCP - I show how to build yours. Which task would you automate first? youtu.be/zbhgn-IBV_I
1
77
What a room - our 19 Nov meetup in DIFC. Share this photo with your reflections and tag Decoding Data Science. Most engagement wins our Evaluations for AI course + a JetBrains product licence. Winner announced 7 Dec. Bootcamp next: 27-29 Nov decodingdatascience.com/boot…
2
1
3
88
Enter: share the photo + your reflections, tag Decoding Data Science, and email your post link + reactions/comments screenshot to [email protected] by 8 PM Dubai, 6 Dec. Free, one entry each. Engagement = reactions + comments; ties go to the earlier post. Prize: course + one eligible JetBrains personal product licence, redeem by 13 Feb 2027.
13
Meet the Top 20 finalists of our 14th Building Agentic AI Application Challenge! Congratulations to every finalist. Judging is under way, with the Top 5 to be announced at our next online meetup. Which project would you like to see demoed? #AgenticAI
1
3
80
Claude Fable 5.2 is out. Anthropic's reported changes: -> up to 30% faster than Fable 5.1 -> stronger coding, longer agent runs -> Mythos variant gated to verified orgs Reported results, not guarantees. Test on your own task first.
1
64
Abu Dhabi put an AI assistant inside its health record system, not beside it. It summarizes a patient's history across 4B+ clinical records. Physicians skip the search and keep the decisions. Enterprise AI pattern: inside the workflow, human in charge. Where would this fit?
47
Stop asking which AI model is best. Ask which model does which job. A strong agent system splits the work: -> a planner reasons -> a worker executes -> a verifier checks Each role can be a different model, sized to its task. Orchestration > model choice.
1
47
One user. One problem. One working demo. That's the strategy behind my AI Accelerator Bootcamp: Clarity -> Prototype -> Evidence. What would prove your AI app works for a real user? decodingdatascience.com/boot…
2
1
43
Alhamdulillah! 🎉 Our AI Residency programme is now CPD certified. 20 masterclasses. 40 CPD learning hours. Building, evaluating and deploying AI applications. Thank you to our learners, mentors and community for being part of this journey. 💙 What will you build next?
1
43
AI news | Nov 11 Claude Fable 5.2 + Mythos 5.2 Meta's Muse Ember for agent tasks Google's TPU 8tx + 8i Max systems Australia plans mandatory AI safety protections Disney + Adobe: custom AI models Which matters most for your work: capability, cost or control?
72
AI agents need boundaries before autonomy. An AWS engineer sets which commands coding agents can run alone and which need approval. A useful lesson as Apple reportedly plans a more agentic Siri. What would you let an agent do without asking?
2
3
140