Course Details
Course Details
What You'll Learn
Topic 1: Introduction to OpenAI Agent Development Kit
Topic 2: Building RAG Agent
Topic 3: Orchestrating Multiple Agents
Topic 4: Model context protocol (MCP)
Topic 5: Fine Tuning LLM Models
Course Info
Minimum Entry Requirement
Knowledge and Skills
- Able to operate using computer functions
- Minimum 3 GCE ‘O’ Levels Passes including English or WPL Level 5 (Average of Reading, Listening, Speaking & Writing Scores)
Attitude
- Positive Learning Attitude
- Enthusiastic Learner
Experience
- Minimum of 1 year of working experience.
Target Age Group: 18-65 years old
Minimum Software/Hardware Requirement
Software:
You will need the following for this course:
- Python 3 – free
- Visual Studio Code – free
- An OpenAI API account – pay-as-you-go API credits are required
Hardware: Windows PC or Mac laptop
Job Roles
Job Roles
- NLP (Natural Language Processing) Engineer
- Machine Learning Engineer
- AI Application Developer
- Chatbot Developer
- AI Research Scientist
- AI Product Manager
- Data Scientist specializing in text data
- Conversational AI Designer
- AI Integration Specialist
- Language Model Trainer
- Content Personalization Engineer
- AI Platform Engineer
- Cognitive Computing Specialist
- AI Technical Consultant
- AI Software Architect
Trainers
Trainers
Dr. Alfred Ang works with the OpenAI Agents SDK as one of the core frameworks in his multi-agent practice, alongside the Google Gemini Agent SDK, Claude Agent SDK, Microsoft AutoGen, CrewAI, n8n, Langflow and Flowise. A full-stack AI developer and Agentic AI solution architect, he designs collaborative, goal-driven agent architectures for enterprise use cases and leads agentic AI adoption that automates customer service, internal workflows, analytics and decision support with large language models, RAG pipelines and multi-agent systems. He serves as Managing Director, CIO, CTO and Chief Instructional Designer of Tertiary Infotech Academy Pte Ltd and is the founder of Tertiary Courses. His GitHub portfolio of more than 300 public repositories contains building blocks this course assembles: n8n-hrportal pairs human-in-the-loop leave approval with an OpenAI HR chatbot protected by input and output guardrails; n8n-ragdemo2 ingests course brochures into Supabase, Pinecone or Qdrant vector databases for a customer-service RAG chatbot; and kacificerpr runs scheduled AI agents with human approvals over email, Asana and Telegram. His machine learning credentials are equally relevant to the fine-tuning topic: he is a Microsoft Certified Data Scientist (DP-100) and AI Engineer Associate (AI-102), an AWS Certified AI Practitioner, and holds a PhD in Electrical Engineering from NUS, an MSc in Electrical and Electronic Engineering from NTU and a Graduate Diploma in Intelligent Robotics Systems from NUS-ISS. He is also CompTIA Security+ certified, which shapes his attention to guardrails, permissions and safe tool access in agent design.
Develop Multi-Agent Systems with OpenAI Agent SDK is taught the way Dr. Ang builds agents himself, starting simple and adding capability only when it earns its place. The first topic introduces the OpenAI Agents SDK, covering agents, instructions, tools, handoffs, guardrails, sessions and tracing, and he spends time on tracing early because seeing what an agent actually did is the fastest way to debug it. The second topic builds a RAG agent, where his vector database work informs practical choices about chunking, embeddings and retrieval quality. Topic three orchestrates multiple agents, splitting work across specialists that hand off tasks to one another, and he explains when a single well-instructed agent is the better design. Topic four introduces the Model Context Protocol (MCP) for connecting agents to external tools and data in a standard way, and the final topic covers fine-tuning LLM models, including how to judge whether fine-tuning, better prompting or retrieval is the right fix for a given problem. Guardrails run through the whole course, reflecting the input and output checks in his own HR chatbot and the governance perspective of his NTU Certificate in AI Ethics and Governance (Senior Professional). As an ACTA-certified trainer and DACE-certified curriculum developer, he keeps the class code-first for developers with Python experience, so learners leave with a working multi-agent application they can extend. To review the agent and RAG projects behind this course, explore his GitHub portfolio at https://github.com/alfredang and connect with him on LinkedIn at https://www.linkedin.com/in/angchewhoe/
Teh Siew Yee is a data science and digital transformation leader with more than two decades of experience driving organisational strategy and designing data ecosystems across Asia Pacific and global markets. Siew Yee brings expertise in predictive modelling, machine learning and deep learning, with hands-on experience in data architecture, engineering and analytics, and has led cross-geographical teams working with industry leaders across the US, UK, China, India, Japan, South Korea and Australia. Siew Yee's experience spans the technology, education, finance, aerospace and eCommerce sectors, and as an ACLP-certified trainer Siew Yee has developed training programmes for staff from the C-suite to the working level.
In this course, Siew Yee draws on that machine learning and data architecture background to guide learners through building multi-agent systems with OpenAI's Agents SDK. Learners start with an introduction to agent development, then build a RAG agent that retrieves knowledge from source documents. They go on to orchestrate multiple specialised agents that hand off tasks to one another, extend agents with the Model Context Protocol (MCP) so they can use external tools, and explore when and how to fine-tune LLM models. Siew Yee's practical, case-based approach helps participants understand how these components fit together, so they can design agent systems suited to their own organisation's needs.
Dwight Nuwan Fonseka is Head of Data Science at Plano Pte. Ltd. and an ACLP-certified trainer with deep expertise in machine learning, deep learning and AI applications. He has extensive hands-on experience developing predictive models and deep learning solutions in Python and R with TensorFlow and Keras, giving him a solid understanding of how models are trained and evaluated. His professional background spans healthcare, finance and customer analytics, and he teaches AI with attention to both its opportunities and its risks. He also holds a Masters in Education from NTU and has eight years of teaching experience.
In Develop Multi-Agent Systems with OpenAI Agent SDK, Dwight brings his Python and model-building background to agent development. He introduces the OpenAI Agents SDK, then guides learners in building a retrieval-augmented generation agent, orchestrating multiple specialised agents that hand off tasks to one another, and working with the Model Context Protocol. His deep learning experience informs the final topic on fine-tuning LLM models, where he helps learners understand what fine-tuning can and cannot achieve. Developers leave able to design, build and evaluate multi-agent systems that use tools and knowledge under guardrails, with a realistic view of their reliability.
Sanjiv Venkatram is an ACTA-certified consultant and trainer with a proven track record in developing business intelligence, data analytics and IoT-led BI solutions. As CEO of Prudentia Consulting, he promotes the active use of the Microsoft platform, and his career includes roles at Microsoft across the US and APAC, Cognizant, Yazaki North America and Oracle. He has delivered BI and analytics projects at scale, including an APAC-wide Power BI dashboard backed by a SQL-based KPI lake and the functional architecture of a business unit P&L reporting tool for a worldwide financial planning and analysis team, with functional expertise in engineering and programme management.
In this course, Sanjiv brings his solution-development and data background to building agent systems with the OpenAI Agent SDK. Learners are introduced to the agent development kit, build a retrieval-augmented generation (RAG) agent, orchestrate multiple agents to work together, and connect agents to tools and data through the Model Context Protocol, before exploring the fine-tuning of LLM models. His experience of designing systems that turn scattered data into dependable business information helps participants think carefully about data sources, structure and quality, which matter greatly in RAG and multi-agent designs. Learners leave with working agents and a practical sense of how to build agent systems that can be maintained.
Review
Write Your Own Review
- Recommended Courses
- Related Courses