Cover Image for Build an Agentic AI System: From Data to Decisions
Cover Image for Build an Agentic AI System: From Data to Decisions
Avatar for Decoding Data Science
UAE's Largest AI & Tech Community
1 Going

Build an Agentic AI System: From Data to Decisions

Zoom
Registration
Welcome! To join the event, please register below.
About Event

​Sunday, 22 November 2026 | 12:00-2:00 PM Dubai time (GST) | Online

​

​An AI answer is easy to generate. The harder part is knowing which data and rules support it.

​

​In this Sunday masterclass, join Mohammad Arshad for a live build of an agentic AI system that connects a language model to structured data and policy documents, then produces answers you can inspect.

​

​THE PROBLEM

​

​It's Friday evening at a busy airport. Delays are stacking up, crews are stretched and costs are rising. The numbers are in a database. The operating rules are in documents. How do you bring them together to understand what happened?

​

​THE SYSTEM

​

​We'll build one AI agent with Gemini. It will query airport data through SQL and retrieve relevant operating rules through RAG, then bring the results together to address three questions:

​

​- What went wrong on Friday?

​- What did it cost?

​- Which operating rules were broken?

​

​The focus is the whole system: the problem, the model, the tools, the data and the checks around the answer.

​

​WHAT WE'LL WALK THROUGH

​

​- Frame the operational question before choosing the tools.

​- Connect the agent to structured data and policy documents.

​- Bring SQL results and retrieved rules into one answer.

​- Inspect evidence and citations instead of trusting a fluent response.

​- Test the answer and handle missing information.

​

​Bring your questions and think about where this pattern could fit your own work. The airport data and rules are synthetic and created for teaching.

​

​Hosted by Mohammad Arshad, founder of Decoding Data Science.

​

​CONTINUE FROM THE LIVE BUILD TO YOUR OWN APPLICATION

​

​The separate AI Accelerator Bootcamp: AI Curious to AI Builders runs online on 27-29 November 2026. It provides a guided route from a problem and target user to a working AI application prototype.

​

​The CPE Framework structures the work:

​Clarity: define the problem, user, scope and success metric.

​Prototype: connect the data, prompts and workflow.

​Evidence: test output quality and improve the weak points.

​

​Enrolment includes three live build sessions, Day 0 setup resources and a dedicated WhatsApp builder group. Masterclass registration and bootcamp enrolment are separate.

​

​Bootcamp details: https://decodingdatascience.com/bootcamp

Avatar for Decoding Data Science
UAE's Largest AI & Tech Community
1 Going