Adaptation and Development
Contextual, memory-based, policy, skill and model adaptation across multiple time scales.
Social AITopic status
Agent Cognition
Reasoning, planning, world models, reflection and metacognition in persistent agents.
Social AIAgent Harness
The runtime architecture that binds models, context, memory, tools, policies and evaluation into governed agency.
Social AIAgent Interoperability: A2A Alongside MCP
Designing auditable task handoffs between independent agents while MCP exposes tools and data to each runtime.
Building and deploying AI applicationsAgentic Workflows and Harnesses
Multi-step model workflows, harness responsibilities, state, memory and reusable workflow patterns.
Building and deploying AI applicationsAI Engineering: Two Directions
How AI for software engineering and software engineering for AI form two cross-cutting perspectives in modern AI engineering.
Start hereAI Suitability
Assessing whether AI belongs in the task and whether errors are detectable, tolerable and recoverable.
Shaping the buildAnti-Patterns
Common failures that cut across AI application, software, agent and shaping work.
Cross-cutting knowledgeArchitecture and Interfaces
Boundaries, contracts, integration and structural decisions that make AI systems changeable.
Software engineering fundamentalsBouw nuttige, betrouwbare AI-systemen
Een moderne leer- en referentiesite voor AI Engineering, georganiseerd rond vier elkaar versterkende vaardigheidsgebieden.
StartpaginaBuild useful, reliable AI systems
A modern learning and reference site for AI Engineering, organised around four mutually reinforcing skill areas.
HomeBuilding and deploying AI applications
Overview of AI application engineering from model capability to reliable production systems.
Building and deploying AI applicationsCo-adaptation and Trust
Negotiated adaptation, calibrated trust, repair, uncertainty and user agency in social AI.
Social AICode Quality and Version Control
Maintainability, readable diffs and reversible change in human and agent-written code.
Software engineering fundamentalsCoding Agent Tool Profiles
Concrete operating principles for Codex, Claude Code and Gemini Code Assist.
Using coding agentsCoding Assistants Versus Coding Agents
The difference between local code help and goal-directed multi-step agent work.
Using coding agentsConfiguration, Permissions and Sandboxing
Model configuration, MCP integrations, execution boundaries, approvals and secrets.
Using coding agentsContext Engineering
Designing the complete information environment available to a model at a particular step.
Building and deploying AI applicationsContext Engineering in Practice
Detailed context design for dynamic agent state, retrieval, memory, trust, token budgets, compression and tool-feedback loops.
Building and deploying AI applicationsContext, Instructions and Planning
Managing finite context, persistent project guidance, skills and adaptive plans.
Using coding agentsCoordination Risks and Concurrency Safety
Cost, context, communication, conflict and stopping-condition risks in multi-agent work.
Using coding agentsData Modelling and Data Engineering
Representing entities, relations, events and lifecycle so AI has meaningful data to work with.
Software engineering fundamentalsDeep Learning
Machine learning using multilayer neural networks, underpinning modern vision, speech, multimodal systems and LLMs.
Building and deploying AI applicationsDelegation, Handoffs and Orchestration
How orchestrators assign, combine and account for work across agents.
Using coding agentsDeterministic-Probabilistic Boundaries
Deciding what belongs in rules or conventional code and what can responsibly use a model.
Shaping the buildDevOps, Deployment and Observability
Build, release and operations practices for observable and recoverable AI software.
Software engineering fundamentalsEmbeddings, Vector Search, Graphs and Semantic Layers
Retrieval building blocks and their trade-offs for semantic, structured and relationship-heavy information.
Building and deploying AI applicationsEvals and Error Analysis
Structured tests and qualitative inspection for improving probabilistic AI-system behaviour.
Building and deploying AI applicationsEvals as the Spine
How evals connect shaping, AI application engineering, software engineering and coding-agent work.
Cross-cutting knowledgeEvaluating Coding Agents
Evidence dimensions for judging agent output, process quality, safety and robustness.
Using coding agentsFoundation Models and LLMs
Foundation models, LLMs, multimodal models and model selection as production engineering choices.
Building and deploying AI applicationsFoundations of Agency
The perception-action loop, state, goals, environment and autonomy as the basis of social agents.
Social AIGlossary
Anchorable definitions for the knowledge base vocabulary.
Cross-cutting knowledgeGovernance and Human Decision Boundaries
Where humans retain authority over risk, irreversible action and release decisions.
Using coding agentsGrounding and RAG
Connecting model output to trusted, task-relevant sources or system state.
Building and deploying AI applicationsHow to Use This Knowledge Base
Use the site for sequential learning, quick reference and cross-area reasoning.
Start hereHuman Oversight and Guardrails
Controls that constrain AI inputs, outputs or actions and keep human authority meaningful.
Building and deploying AI applicationsLLM Learning
How large language models learn through pretraining, post-training and runtime adaptation, and what that means for AI system design.
Building and deploying AI applicationsLLM Training Example: Git
A worked example of how pretraining, post-training, trajectories, feedback and an agent harness produce procedural Git competence.
Building and deploying AI applicationsLong-Lived Identity
Identity continuity, layered stability and model migration for agents that persist for years.
Social AIMachine Learning
Methods that learn patterns from data instead of expressing every rule explicitly.
Building and deploying AI applicationsMemory Systems
Working, episodic, semantic, procedural, relational and self-memory, plus lifecycle governance.
Social AIModel, Harness, Tools and Workspace
The practical components that determine what a coding agent can do.
Using coding agentsMulti-Agent Patterns
Fan-out/fan-in, pipelines, reviewer patterns and debate patterns.
Using coding agentsMultiple Agents and Subagents
When separate agents help and when a single agent is the better engineering choice.
Using coding agentsOutcomes and Learning Loops
Outcome metrics, guardrail metrics, build-measure-learn loops and decision records.
Shaping the buildPatterns
Reusable cross-area patterns for practical AI Engineering.
Cross-cutting knowledgePrinciples for Effective Use
Operational principles for getting useful, reviewable and reversible work from coding agents.
Using coding agentsProblem Framing and Stakeholder Analysis
Defining the situation, affected people, desired change and competing interests before selecting an AI solution.
Shaping the buildProduct Hypotheses and Prototypes
Testing falsifiable product hypotheses with prototypes and vertical slices.
Shaping the buildProduction Operations, Observability, Cost and Latency
Running AI systems under real deployment, tracing, reliability, cost and performance constraints.
Building and deploying AI applicationsPrompt Engineering
Designing instructions and examples that elicit useful model behaviour.
Building and deploying AI applicationsQuality Model for AI Software
A four-layer definition of done for AI software.
Cross-cutting knowledgeReal World Examples
Worked examples that show how the AI Engineering approach applies to realistic product and operations settings.
Real world examplesRequirements and Design
Requirements engineering, computational thinking and design for AI-supported systems.
Software engineering fundamentalsRetail Logistics Intelligent Agent
A worked AI Engineering example for an agent that advises on fulfillment, warehouse operations and distribution planning in a retailer.
Real world examplesRisk and Responsible AI
Assessing hazards and designing accountable AI systems with meaningful human oversight.
Shaping the buildSecurity, Privacy and Accessibility
Security, privacy and inclusive design constraints for AI-supported applications.
Software engineering fundamentalsShaping Canvas
A worksheet for turning an AI idea into a valuable, feasible, responsible and testable build.
Shaping the buildShaping the build
Overview of shaping valuable, feasible, responsible and testable AI interventions.
Shaping the buildSkills and Goals
Reusable procedural competence, goal typing and skill learning without opaque self-modification.
Social AISocial AI
Long-lived adaptive social AI agents, from agency foundations to memory, identity, co-adaptation, trust and evaluation.
Social AISocial State and Relationships
User models, self models, relationship models, theory of mind, social norms and relational memory.
Social AISoftware engineering fundamentals
Overview of the software fundamentals that make AI systems understandable, testable and operable.
Software engineering fundamentalsSources
Source links and maintenance note for the AI Engineering Knowledge Base.
Cross-cutting knowledgeSpecifications, Quality Attributes and Acceptance Criteria
Turning desired behaviour, constraints and quality into verifiable conditions.
Shaping the buildSuggested Learning Paths
Three ordered paths for students, builders and agentic development practice.
Start hereTask Briefs and Examples
A reusable task brief structure for bounded, verifiable agent work.
Using coding agentsTechnical Debt, Documentation and Teamwork
Managing future cost, decision records, collaboration and review in fast-moving AI projects.
Software engineering fundamentalsTesting and Testability
Traditional tests for deterministic behaviour and evals for probabilistic behaviour.
Software engineering fundamentalsThe Agent Loop
Observe, reason, act, inspect, update and stop or continue.
Using coding agentsThe AI Engineer
The professional role, attitude, competencies and oversight responsibilities of the AI engineer.
Start hereThe Four-Area Skills Map
The four mutually reinforcing areas that organise this knowledge base.
Start hereThe Ten-Step Working Method
A readable process for directing coding agents from task framing through shipping and observation.
Using coding agentsTool Calling and MCP
Allowing models to request controlled operations through validated tools and standard interfaces.
Building and deploying AI applicationsTopic status
Browse topics by their publication and development status.
Cross-cutting knowledgeTrustworthy Agency and Evaluation
Alignment, bounded autonomy, uncertainty, privacy, safety, governance and multi-timescale evaluation.
Social AIUser and Workflow Research
Understanding tasks, handoffs, exceptions and user progress before changing a workflow.
Shaping the buildUsing coding agents
Overview of coding agents, their working environment, multi-agent orchestration and evidence-based completion.
Using coding agentsValue, Feasibility, Viability and Responsibility
Four tests for deciding whether a shaped AI intervention deserves further investment.
Shaping the buildWhat is AI Engineering?
AI Engineering is the discipline of designing, building, evaluating, deploying and improving software systems in which AI models perform part of the system work.
Start here