Agentic AI Developer Program

Course Description

The Agentic AI Developer Program is a hands-on, project-based course for professionals who want to design, build, evaluate and deploy AI agents that do real work. Working with Python, local and open-source language models and real business data, you will build agents that solve genuine organisational challenges, not classroom exercises.

You will learn to connect language models with vector databases, APIs and business systems such as CRMs and relational databases. The emphasis throughout is on what makes agents trustworthy in practice: tool orchestration, validation, human-in-the-loop approval and reliability controls. By the end, your agent will be live and accessible through a web interface, Telegram or an API.

Program Schedule and Fees

The program includes 15 hours of live, instructor-led training, delivered in person at Dubai Knowledge Park or live online, with guided capstone development. Each cohort is limited to 12 participants to ensure personal attention.

Cohort Delivery Location Dates Schedule Fee
December Weekend Cohort In person or live online Dubai Knowledge Park 05, 06, 12, 13, 19, 20 December 2026 Saturday and Sunday
3:00 PM–5:30 PM GST		 USD 1,500
VAT inclusive

Private corporate cohorts and group rates for three or more participants are available on request.

Ask about upcoming cohorts or request a corporate quotation

Who This Program Is For

This program is for people who are ready to move beyond using AI tools and start building AI systems of their own.

  • Developers and software professionals
  • Data analysts, data scientists and data engineers
  • IT and technical professionals working with automation, APIs or business systems
  • Entrepreneurs and technically minded professionals who want to build AI-agent applications

Prerequisite: Basic Python familiarity is recommended; advanced software-engineering experience is not required. All work is done in guided Jupyter notebooks in a hosted environment, so you only need to bring a laptop. If you would like to strengthen your Python first, start with Python Programming Foundations.

A Natural Next Step for AI Professional Program Alumni

Completed Innosoft Gulf’s AI Professional Program? The Agentic AI Developer Program builds directly on your Python and AI foundations and takes you into hands-on agent development.

Alumni are eligible for a special registration offer. Contact us for details.

Build Your Own Working AI Agent

Every participant leaves with a working agent. Bring your own idea or choose one of the project directions below, and the instructor will help you shape it into a realistic scope you can build and demonstrate within the program.

  • Lead Generation and Sales Follow-Up Agent
  • Real Estate Enquiry and Property Agent
  • Productivity and Administrative Agent
  • Document Processing and Compliance Agent
  • Financial Market Intelligence and Monitoring Agent
  • Business Research and Management-Reporting Agent

Your finished agent can be presented through Telegram, a web interface or an API. Technically suitable projects can be deployed on Innosoft Gulf’s Kubernetes infrastructure, with continued production hosting and long-term support available separately.

Development approach: Define the agent project → build the workflow → make it reliable and controlled → deploy and demonstrate.

Learning Outcomes

By the end of the program, you will be able to:

  • Distinguish chatbots, automated workflows and agentic AI systems.
  • Build controlled agent workflows using Python and guided Jupyter notebooks.
  • Apply reflection, critique and evaluation patterns to improve agent outputs.
  • Create and connect approved tools, structured outputs and business functions.
  • Build document-intelligence and RAG workflows using local embeddings and vector search.
  • Design human-approval boundaries, validation and recovery mechanisms for agent actions.
  • Assess agent reliability, privacy, safety and operational limitations.
  • Develop, deploy and demonstrate a working AI agent.

Ready to Build Your Own AI Agent?

Apply today and tell us about the agent you would like to build.

Apply Now Contact Innosoft Gulf

Program Curriculum

Eight modules take you from the foundations of agentic workflows through tools, RAG, human controls and reliability to building and deploying your own agent.

Module 1

From Language Models to Agentic Workflows

Key Topics

  • Chatbots, workflows and agents: the practical differences
  • Degrees of autonomy and the right level of human oversight
  • Task decomposition, state and controlled information sources
  • Running a local small language model with Ollama
Practical lab: Build a controlled research agent.

Module 2

Reflection, Critique and Output Improvement

Key Topics

  • Direct generation versus review-and-revise workflows
  • Generator, critic and reviser roles
  • Business rubrics and measurable quality criteria
  • Automated evaluation checks and human feedback
Practical lab: Build a reflection workflow that critiques and improves its own output.

Module 3

Tool Use, Structured Outputs and Safe Refusal

Key Topics

  • Tool and function calling for business applications
  • Creating approved Python tools with defined inputs and outputs
  • Structured extraction and routing requests to the right tool
  • Model-only answers versus tool-grounded responses
  • Logging tool actions and refusing when reliable evidence is unavailable
Practical lab: Build an agent that selects and uses approved business tools.

Module 4

Document Intelligence and Multimodal RAG

Key Topics

  • PDF text extraction and page rendering
  • Analysing tables, charts and visual page content
  • Chunking, local embeddings and vector search with Weaviate
  • Grounded retrieval and source-aware answers
  • Local vision and language models for private document workflows
Practical lab: Build a multimodal document question-answering workflow.

Module 5

Controlled Workflows and Human Approval

Key Topics

  • Deterministic business rules versus model interpretation
  • Read-only actions versus actions that change business data
  • Clarification, availability checks and pending actions
  • Human approval before create, update or delete operations
  • Case study: a calendar agent built with Telegram, Google Calendar and local inference
Practical lab: Design a multi-step workflow with a human-approval boundary.

Module 6

Reliability, Validation and Recovery

Key Topics

  • Planning an action before executing it
  • Validating generated code and structured actions
  • Repairing invalid outputs and restricting execution
  • Database updates, verification and automatic rollback
  • Visible fallbacks and audit-friendly logs
Practical lab: Build a reliable, locally run customer-service agent.

Module 7

Capstone Project Development

Key Topics

  • Defining a realistic business or organisational problem
  • Designing the agent architecture, approved tools and boundaries
  • Combining tool use with at least two other agentic patterns
  • Adding safety checks, logging and evaluation tests
  • Instructor review and guided development
Project activity: Develop and test your capstone AI agent.

Module 8

Deployment, Evaluation and Project Demonstration

Key Topics

  • Integration through Telegram, an API or a web interface
  • Local, cloud and hybrid deployment trade-offs
  • Containerisation and Kubernetes deployment
  • Logging, health checks, privacy and access boundaries
  • Limitations, required human oversight and an implementation roadmap
Final presentation: Demonstrate your working agent and walk through its architecture, data sources, tools, safeguards, limitations and deployment approach.

Learning Experience and Participant Resources

You learn by building: live instructor-led sessions, guided Jupyter notebooks, practical labs and a capstone project, all supported by a ready-to-use technical environment.

  • Open-source AI models: local language and vision models through Ollama.
  • Databases and project data: vector search, relational data, documents, PDFs, tables, images and approved datasets.
  • Hosted development environment: managed JupyterHub, guided notebooks, browser-based coding workspaces and preinstalled Python libraries.
  • Deployment platform: containerised hosting for technically suitable capstone projects.
  • Access throughout the program and the agreed project-development period, with extended hosting available separately.
  • Instructor review, supporting materials and ongoing project guidance.
  • Certificate of Completion: awarded on meeting attendance, lab and final project requirements. KHDA attestation is available for eligible certificates.
  • Optional professional certification: Innosoft Gulf is developing a separate Innosoft Certified Agentic AI Professional assessment. Details will be announced soon.

What Participants Say

“If you want to master AI agents, this is the real deal. The progression from core basics such as RAG to complex workflows is incredibly well structured, and Innosoft provides a high-performance platform and JupyterLab environment for hands-on practice.”

Eureka Ilunga

Software Engineer

Feedback on the program’s previous edition, AI Agents for Business Applications.

“Great courses with plenty of real-world examples and hands-on labs. The Agentic AI training was engaging, practical and well structured. Ahmed explains challenging concepts clearly and creates a positive learning atmosphere.”

Paulo Araujo

Data Science Professor

“They are truly passionate about the subject, and their energy is contagious. They create a supportive learning environment that encourages participation and critical thinking.”

Maggie Chung

Sales and Marketing Professional

Testimonials are drawn from participant feedback on Innosoft Gulf training programs. Some have been shortened for presentation without changing their meaning.

Start Building Your AI Agent

Join a small cohort and leave with a working AI agent, built with Python, approved tools and expert guidance.

Apply Now Contact Innosoft Gulf
Scroll to Top