Course Outline
Foundations of Architecture for AI Systems
- Layers, boundaries, and the principle of separation of concerns
- Applying ports and adapters to AI components
- Defining quality attribute scenarios for AI features
- Using architecture decision records for governance
Evolving Existing Applications
- Incremental modernisation rather than full rewrites
- Positioning AI services alongside existing monoliths
- Strangler pattern and anti-corruption layer strategies
- Integration seams for existing APIs and data structures
LLM Integration Architecture Patterns
- The model gateway as an abstraction boundary
- Synchronous, asynchronous, and streaming integration methods
- Fallback strategies, degradation handling, and provider failover
- Managing prompt and model configuration as versioned assets
Contracts, APIs, and Data Exchange
- Structured output schemas and validation techniques
- Handling idempotency, retries, and rate limiting
- Latency budgets and timeout orchestration
- Cost per request as a key architectural constraint
End-to-End RAG Architecture
- Ingestion, chunking, and embedding pipelines
- Stages for indexing, retrieval, reranking, and generation
- Positioning retrieval relative to application services
- Evaluating trade-offs and failure modes at each stage
Vector Databases and Retrieval Design
- Comparing managed versus self-hosted vector stores
- Dense, sparse, and hybrid search methodologies
- Metadata filtering and tenant isolation
- Re-embedding strategies for model changes
Retrieval Quality, Grounding, and Evaluation
- Building evaluation sets from real-world queries
- Metrics for retrieval and answer quality
- Citation, grounding, and containing hallucinations
- Implementing feedback loops into the index
Agent Architectures
- The agent loop and tool invocation mechanics
- Deterministic workflows versus autonomous planners
- Identifying cases where an agent is not the right solution
- Scoping agent responsibilities and state management
Multi-Agent Systems and Orchestration
- Supervisor and orchestrator topologies
- Handoff protocols and message contracts
- Shared memory and context management
- Framework comparison based on architectural merits
Service Communication and Tool Protocols
- Event-driven and queue-based coordination
- Exposing tools via standard protocols
- Sagas, compensation, and long-running transactions
- Observability across service hops
AI Agents in the Development Workflow
- Incorporating coding agents into the delivery pipeline
- Spec-driven development and review gates
- Test generation and automated quality checks
- Guardrails that maintain architectural control
Governance and Software Quality
- Design reviews for AI features
- Managing model and prompt changes
- Regression evaluation as a pipeline stage
- Defining ownership, policy, and approval paths
Security, Privacy, and Data Governance
- Defending against prompt injection with layered security
- Managing secrets, access control, and least privilege
- Data residency, personal data, and retention policies
- Auditability and human-in-the-loop controls
Scalability, Cost, and Observability
- Tracing across agent chains
- Token usage and cost telemetry
- Caching and latency optimization
- Service levels and incident response for AI services
Capstone: Architecture for an AI-Powered Solution
- Context, component, and deployment views
- Designing retrieval and agent orchestration
- Planning for security, governance, and cost
- Peer review, critique, and migration roadmap
Requirements
- Experience in designing and maintaining production-grade software systems
- Working knowledge of REST APIs and service integration
- Familiarity with relational databases and data modelling
- Basic exposure to LLM-based tools or APIs
Target Audience
- Software architects
- Technical leads and senior engineers
- Engineering managers accountable for architectural decisions
Testimonials (1)
That i gained a knowledge regarding streamlit library from python and for sure i'll try to use it to improve applications in my team which are made in R shiny