AI SOLUTION ARCHITECT

Explore how this course is designed to take you from foundational concepts to production-ready AI systems. Understand the evolving AI landscape — from traditional machine learning to generative and agentic systems — and how these capabilities are reshaping architecture roles. 

• Meet your instructor
• Course structure, grading, & expectations
• Assignments & final project overview
• The AI landscape
• Bridging architecture and AI engineering
• AI-assisted development for architects

Define the role and responsibilities of an AI Solution Architect in the era of Generative AI. Learn to Distinguish between traditional ML, Generative AI, and Agentic AI use cases and formulate an AI solution proposal from a business problem.

• Role & responsibilities 
• Case Study: AI technologies & applications
• Formulating AI solutions

Learn to source, prepare, and validate data for modern AI systems. Work with structured, unstructured, and multimodal data, and apply preprocessing techniques optimised for LLMs and RAG pipelines. 

• Data sources & collection strategies
• Document AI and parsing tools 
• Data preprocessing techniques
• Data quality assessment
• Demo: End-to-end data preprocessing pipeline 

Assignment #1: Build a preprocessing pipeline that cleans, chunks, and embeds text data, and evaluates retrieval performance.

Explore how machine learning evolved into modern Generative AI. Understand core ML paradigms, neural networks, and transformers, and learn when to apply traditional ML versus GenAI approaches. 

• Machine learning pipeline & workflow
• Neural network essentials
• The path to Generative AI
• NLP in the age of LLMs
• Demo: ML vs LLM comparison

Dive into modern LLM architectures and learn to use them effectively. Apply prompt engineering, design structured outputs, and build RAG systems that ground models in real data. 

• LLM architecture
• Prompt engineering
• Retrieval-augmented generation 
• Case Study: Industry use cases for Generative AI
• Demo: Building a simple RAG pipeline end-to-end 

Bonus Assignment: Implement a basic RAG pipeline using a provided dataset.

Design data pipelines that power AI and GenAI systems. Learn to structure workflows for training, inference, and retrieval, and work with vector databases for semantic search. 

• Data engineering lifecycle for AI
• Data pipeline architectures
• Big data & AI platforms
• Vector databases & embeddings infrastructure
• Demo: Building an ingestion pipeline

Assignment #2: Build a document ingestion and embedding pipeline using a vector database.

Design advanced RAG systems that go beyond basic retrieval. Apply techniques like hybrid search, re-ranking, and query transformation, and evaluate system performance using real metrics. 

• Advanced RAG patterns
• Agentic RAG
• RAG evaluation
• Production RAG considerations 
• Demo: Build an advanced RAG system

Assignment #3: Enhance your RAG system with advanced techniques and evaluate improvements.

Design AI systems that can reason, plan, and act. Learn core agent patterns and build multi-agent systems that collaborate to solve complex tasks. 

• AI agent fundamentals
• Agentic design patterns
• Multi-agent systems
• Computer use & browser agents
• Voice & realtime agents
• Agent frameworks comparison & agentic protocols

Assignment #4: Build a multi-agent system with specialised roles working together.

Compare AI services across cloud platforms and design scalable, secure architectures. Learn to deploy AI systems while meeting cost, performance, and data residency requirements. 

• Cloud architecture patterns for AI
• Sovereign AI & data residency 
• Training and deploying models on cloud
• Cloud security for AI workloads
• Demo: Deploying LLM-powered application on one cloud platform

Assignment #5: Deploy an AI endpoint and benchmark performance and cost.

Optimise AI systems for scale. Apply techniques to reduce latency and cost, and design architectures that balance performance, quality, and efficiency. 

• Scaling AI solutions
• AI gateways, LLM proxies, specialised inference providers
• Cost optimisation & performance benchmarking
• Small language models & edge deployment
• Demo: Scaling and optimising a GenAI inference pipeline
• Lab: Take a baseline LLM inference setup and apply at least 3 optimisation techniques

Apply MLOps principles to production AI systems. Manage models, prompts, and pipelines while implementing monitoring, deployment, and security best practices. 

• MLOps fundamentals
• GenAIOps: The evolution
• Deployment strategies
• Security & risk mitigation
• Observability for AI systems

Build reliable agent systems for production. Design for failure handling, human oversight, and long-running workflows while managing drift, cost, and system resilience. 

• Agent reliability patterns
• Human-in-the-loop architecture
• Multi-LLM resilience & fallback strategies 
• GenAI-specific drift detection 
• Knowledge base maintenance for RAG
• Agentic compliance & audit 

Assignment #6: Set up monitoring and alerting for an AI system.

Control AI system behavior through context, memory, and evaluation. Optimise inputs, design memory systems, and implement continuous evaluation pipelines for production use. 

• Context engineering & agent memory architectures
• Model routing & mixture-of-models in production
• Fine-tuning decision framework 
• Continuous evaluation in production
• Demo: End-to-end evaluation pipeline 
• Lab: Build an automated evaluation pipeline for your RAG or agent system

Communicate AI systems clearly to technical and business stakeholders. Present architectures, quantify impact, and build trust through structured, transparent communication. 

• Translating AI solutions for business
• Presentation frameworks for AI solutions
• Stakeholder communication
• Demo: Presenting an AI Solution Architecture that includes GenAI components

Assignment #7: Create a 10-minute executive presentation of your solution

Identify ethical risks in generative and agentic AI systems. Apply responsible AI frameworks and design solutions that address bias, security, and regulatory requirements. 

• Ethical considerations in Generative AI
• Bias and fairness
• Responsible AI frameworks
• Regulatory landscape
• Security as an ethical imperative
• Lab: Conduct an ethical audit and define mitigation strategies 

Present a complete AI solution from concept to implementation. Demonstrate your system, explain your decisions, and evaluate production readiness. 

• Capstone presentations and demos
• Architecture walkthroughs
• Production readiness evaluation
• Responsible AI considerations
• Peer feedback and review

Explore career paths in AI architecture and plan your next steps. Understand emerging trends and build a strategy for continued growth in the field.

• Career paths for AI solution architects
• Certifications & continued learning
• Emerging trends
• Course wrap-up