AI Application Development Mastery

20-Part Foundation Track

Foundation Track — SDK-Agnostic

A complete, framework-agnostic curriculum covering every layer of modern AI application development. Concepts, patterns, and architectures apply regardless of which SDK or vendor you use. Each part is independently readable.

Part 1

Foundations & Evolution of AI Apps

35 min read

HistoryTransformersLLM Revolution

Trace the evolution from ELIZA and expert systems through the transformer revolution to modern agent-based architectures. Includes a comprehensive framework landscape comparison.

Read Article →

Part 2

LLM Fundamentals for Developers

30 min read

TokenizationContext WindowSampling

Master how LLMs work from a developer’s perspective — tokenization, context windows, temperature, top-p, streaming, and building your first LLM-powered app.

Read Article →

Part 3

Prompt Engineering Mastery

30 min read

Zero-ShotChain-of-ThoughtReAct

Zero-shot, few-shot, chain-of-thought, ReAct, structured outputs, and prompt optimisation techniques that work across any LLM API.

Read Article →

Part 4

Orchestration & Chaining Patterns

35 min read

ChainingPipelinesRunnables

Understand orchestration and chaining concepts — sequential pipelines, parallel execution, conditional branching, output parsers, and the Runnable interface pattern across any framework.

Read Article →

Part 5

Retrieval-Augmented Generation

35 min read

EmbeddingsVector DBRAG

Understand RAG architecture from embeddings to vector databases, chunking strategies, retrieval techniques, re-ranking, and building production RAG pipelines.

Read Article →

Part 6

Memory & Context Engineering

30 min read

Buffer MemorySummaryRe-ranking

Design memory systems for LLM applications — buffer memory, summary memory, vector memory, conversation window management, and context compression strategies.

Read Article →

Part 7

Agents — Core of Modern AI Apps

35 min read

ReActTool-CallingAgents

Build autonomous agents using the ReAct pattern, tool-calling primitives, planner-executor architectures, and self-reflective agent loops.

Read Article →

Part 8

Stateful Agent Workflows

35 min read

State MachinesGraphsWorkflows

Design stateful multi-step agent workflows using graph-based execution models — nodes, edges, conditional routing, cycle handling, and state persistence patterns.

Read Article →

Part 9

Deep Agents & Autonomous Systems

35 min read

Deep AgentsPlanningReflection

Implement deep agents capable of multi-step reasoning, self-reflection, dynamic planning, tool-use sandboxes, and adaptive execution across long-horizon tasks.

Read Article →

Part 10

Multi-Agent Systems

30 min read

SupervisorSwarmCollaboration

Design multi-agent systems using supervisor, swarm, debate, and role-based collaboration patterns for complex task decomposition and parallel problem-solving.

Read Article →

Part 11

AI Application Design Patterns

30 min read

Design PatternsArchitectureAgent Loops

Apply proven design patterns for AI applications — RAG pipelines, chat with memory, workflow automation, agent loops, and hybrid symbolic-neural systems.

Read Article →

Part 12

AI App Dev Ecosystem

35 min read

FrameworksEcosystemSelection Guide

Navigate the AI framework ecosystem — compare LangChain, LlamaIndex, Haystack, AutoGen, CrewAI, and others. Understand how to choose and combine tools for your use case.

Read Article →

Part 13

MCP Foundations & Architecture

35 min read

MCPProtocolTools

Understand the Model Context Protocol (MCP) — its Host/Client/Server architecture, primitive types (tools, resources, prompts), transport layers, and security model.

Read Article →

Part 14

MCP in Production

30 min read

MCP ServersIntegrationScaling

Build production MCP servers, integrate with orchestration frameworks, implement authentication and rate limiting, and scale MCP-based AI systems.

Read Article →

Part 15

Evaluation & LLMOps

35 min read

EvaluationTracingLLMOps

Evaluate AI application quality with automated metrics, distributed tracing, experiment tracking, A/B testing, and LLMOps workflows for continuous improvement.

Read Article →

Part 16

Production AI Systems

30 min read

ProductionAsyncScaling

Deploy AI apps to production with async APIs, message queues, response caching, streaming endpoints, load balancing, and horizontal scaling strategies.

Read Article →

Part 17

Safety, Guardrails & Reliability

30 min read

SafetyGuardrailsPrompt Injection

Implement input filtering, output validation, hallucination mitigation, prompt injection defence, and reliability patterns for production AI systems.

Read Article →

Part 18

Advanced Topics

30 min read

Fine-tuningTool LearningHybrid AI

Explore fine-tuning for domain adaptation, tool learning, hybrid LLM+symbolic reasoning, and cutting-edge techniques shaping the next generation of AI applications.

Read Article →

Part 19

Building Real AI Applications

40 min read

ChatbotDocument QAFull-Stack

Apply everything to real applications: a customer support chatbot, document QA system, coding assistant, and a full-stack AI product from idea to deployment.

Read Article →

Part 20

Future of AI Applications

25 min read

Autonomous AgentsMultimodalFuture

Look ahead at autonomous agents, self-improving systems, multi-modal AI, AI operating systems, and the architectural patterns that will define the next decade.

Read Article →

LangChain SDK Track — 8 Parts

LangChain & LangGraph SDK Deep Dive

A focused, hands-on track covering the LangChain ecosystem from core to advanced — LCEL chains, retrievers, memory, agents, LangGraph workflows, LangSmith observability, MCP tool interoperability, and the Deep Agents harness framework.

LC Part 1

Setup, Chains & LCEL

40 min read

LCELRunnablesOutput Parsers

Install and configure LangChain. Master LCEL (LangChain Expression Language), the Runnable interface, chains, output parsers, and prompt templates with full code examples.

Read Article →

LC Part 2

RAG & Retrievers

45 min read

VectorstoresRetrieversRAG Pipeline

Build RAG pipelines with LangChain: document loaders, text splitters, vectorstore integrations (Chroma, Pinecone, FAISS), retrievers, and the full RetrievalQA chain.

Read Article →

LC Part 3

Memory & State Management

35 min read

ConversationMemoryStatePersistence

LangChain memory implementations: ConversationBufferMemory, SummaryMemory, VectorStoreRetrieverMemory, and building stateful conversation applications.

Read Article →

LC Part 4

Agents & Tools

45 min read

AgentExecutorToolsToolkits

Build LangChain agents with AgentExecutor, create custom tools and toolkits, use built-in tools (Search, Python REPL, SQL), and implement structured tool output parsing.

Read Article →

LC Part 5

LangGraph — Stateful Workflows

50 min read

StateGraphEdgesCheckpointers

Master LangGraph: StateGraph, nodes, conditional edges, MemorySaver checkpointers, human-in-the-loop interrupts, and building multi-agent workflows with persistence.

Read Article →

LC Part 6

LangSmith — Tracing & LLMOps

40 min read

LangSmithTracingEvaluation

Instrument LangChain apps with LangSmith for full trace visibility, run automated evaluations, create datasets, and integrate LangSmith into CI/CD for continuous LLMOps.

Read Article →

LC Part 7

MCP — Model Context Protocol

45 min read

MCPFastMCPTool Servers

Build and connect MCP tool servers with LangChain agents using langchain-mcp-adapters: FastMCP, MultiServerMCPClient, transports, stateful sessions, and production patterns.

Read Article →

LC Part 8

Deep Agents & Harness Framework

55 min read

Deep AgentsHarnessSubagents

The batteries-included agent harness: create_deep_agent(), virtual filesystem, context engineering, subagent delegation, skills, memory, middleware, and production deployment.

Read Article →

OpenAI SDK Track — 20 Parts

OpenAI SDK Deep Dive

A comprehensive, 20-part track covering the OpenAI platform from account setup through production operations — Responses API, structured outputs, function calling, multimodal (vision, audio), embeddings, Agents SDK, Realtime API, reasoning systems, prompt engineering, safety, production engineering, observability, enterprise compliance, advanced architectures, and capstone projects.

OA Part 1

Platform & SDK Setup

35 min read

API KeysRate LimitsError Handling

OpenAI platform setup: organisations, projects, billing, rate limits, SDK installation (Python & TypeScript), sync/async client architecture, and production error-handling patterns.

Read Article →

OA Part 2

Responses API & Text Generation

40 min read

Responses APIStreamingReasoning Models

Master the Responses API: messages, system instructions, generation controls (temperature, top_p, stop), SSE streaming, o-series reasoning models, and multi-turn conversation patterns.

Read Article →

OA Part 3

Structured Outputs & Code Generation

35 min read

JSON SchemaPydanticCode Gen

JSON mode vs structured outputs with json_schema enforcement, Pydantic model integration, schema-first design, code generation patterns, and deterministic output strategies.

Read Article →

OA Part 4

Function Calling & Tools

40 min read

Tool DefinitionsParallel CallsOrchestration

Tool definitions with JSON Schema, parallel tool calling, forced vs auto tool_choice, multi-step tool orchestration, idempotency, retries, and production timeout patterns.

Read Article →

OA Part 5

Multimodal — Vision & Images

40 min read

Vision APIDALL-E 3GPT-Image-1

Vision API for image understanding, OCR, and diagram reasoning; DALL-E 3 and GPT-Image-1 for generation; image editing; multi-image inputs; and production image pipelines.

Read Article →

OA Part 6

Multimodal — Audio, Speech & Video

35 min read

WhisperTTSSora

Whisper STT for transcription and translation, TTS API with voice selection and streaming, audio preprocessing, Sora video generation, and combining audio-visual modalities.

Read Article →

OA Part 7

Embeddings & File Search

40 min read

text-embedding-3Vector StoresFile Search

Embeddings API with text-embedding-3 models, OpenAI vector stores, the File Search built-in tool, native RAG with OpenAI’s retrieval stack, and hybrid search patterns.

Read Article →

OA Part 8

Agents SDK

50 min read

Agent DefinitionsHandoffsGuardrails

OpenAI Agents SDK: agent definitions (instructions, tools, handoffs), Runner lifecycle, multi-agent orchestration, input/output guardrails, tracing, voice agents, and Agent Builder.

Read Article →

OA Part 9

Realtime API

40 min read

WebRTCVADSession Management

WebRTC vs WebSocket vs SIP transports, session lifecycle management, voice activity detection, realtime function calling, handling interruptions, and cost management strategies.

Read Article →

OA Part 10

Fine-Tuning, Eval & Production

50 min read

Fine-TuningEvaluation APIBatch API

SFT workflow, vision and reinforcement fine-tuning, Evaluation API, prompt optimizer, Batch API for cost savings, prompt caching, Moderation API, and enterprise RBAC/audit patterns.

Read Article →

OA Part 11

Reasoning Systems

35 min read

GPT-5.5Reasoning EffortMulti-Step

GPT-5.5 reasoning models, reasoning effort parameter (low to xhigh), reasoning items in multi-turn conversations, tool calling with reasoning, and multi-step decomposition patterns.

Read Article →

OA Part 12

Context & State Management

32 min read

Prompt CachingConversations APIState

Prompt caching with store: true, previous_response_id for chained conversations, Conversations API for persistent state, context window management, and multi-turn best practices.

Read Article →

OA Part 13

Prompt Engineering

38 min read

System InstructionsFew-ShotSecurity

System instructions architecture, few-shot prompting, chain-of-thought patterns, output format control, prompt injection defense, advanced patterns, and prompt testing workflows.

Read Article →

OA Part 14

Safety & Moderation

34 min read

Moderation APIGuardrailsRed Teaming

Moderation API, content filtering strategies, guardrail design, prompt injection defense, PII detection, abuse prevention, and safety testing with red-teaming methodologies.

Read Article →

OA Part 15

Production Engineering

36 min read

Batch APIRate LimitingCost Optimization

Architecture patterns, rate limit engineering, Batch API for 50% savings, caching strategies, deployment patterns, horizontal scaling, and intelligent model routing for cost control.

Read Article →

OA Part 16

Observability & Monitoring

30 min read

TracesOpenTelemetryCost Monitoring

OpenAI Dashboard traces, custom structured logging, metrics and alerting, OpenTelemetry integration, cost monitoring dashboards, quality tracking, and incident response runbooks.

Read Article →

OA Part 17

Enterprise & Compliance

28 min read

Admin APIRBACData Privacy

Organization and Admin API, project isolation, access control and RBAC, zero data retention (ZDR), audit and governance, multi-region with Azure OpenAI, and enterprise security.

Read Article →

OA Part 18

Advanced Architectures

40 min read

Multi-AgentAgentic RAGHuman-in-Loop

Multi-agent systems, agentic RAG with query routing, plan-and-execute patterns, human-in-the-loop workflows, autonomous agents, event-driven architectures, and hybrid designs.

Read Article →

OA Part 19

Migration & Legacy Integration

30 min read

Chat→ResponsesModel UpgradeMulti-Provider

Chat Completions to Responses API migration, model upgrade strategies, SDK version upgrades, legacy system adapters, multi-provider fallback, and eval-driven migration testing.

Read Article →

OA Part 20

Capstone Projects

42 min read

Full-Stack AIPortfolio4 Projects

Four end-to-end projects: AI Document Processor, Multi-Agent Customer Service, Realtime Voice Assistant, and Autonomous Research Agent — integrating all 19 prior parts.

Read Article →

23-Part Anthropic SDK Track

Anthropic Claude SDK & CCA Certification

Complete preparation for the Certified Claude Agent (CCA) exam — covering all 5 domains: Platform & SDK (30%), MCP & Tools (20%), Agent Configuration (20%), Prompt Engineering (15%), and Safety & Administration (15%).

Part 1

Platform & SDK Setup

35 min read

SDK InstallAuthCCA Domain 1

SDK installation, API key authentication, first API call, and model selection strategy for the CCA exam.

Read Article →

Part 2

Messages API & Content Blocks

40 min read

MessagesStreamingMultimodal

Messages API structure, content block types, streaming patterns, multimodal inputs, and system prompts.

Read Article →

Part 3

Agentic Loops & Task Decomposition

40 min read

Agentic Looptool_usestop_reason

The tool_use agentic loop, stop_reason handling, task decomposition, and multi-step agent patterns.

Read Article →

Part 4

Multi-Agent Orchestration

35 min read

Multi-AgentDelegationPatterns

Multi-agent patterns: orchestrator-worker, delegation, parallel execution, and agent communication.

Read Article →

Part 5

Hooks, Sessions & State

35 min read

HooksSessionsState

Hook lifecycle (PreToolCall, PostToolCall, Notification), session persistence, and state management.

Read Article →

Part 6

Tool Interface Design

35 min read

Tool SchemaJSON SchemaPatterns

Tool schema design, JSON Schema validation, descriptions for LLM comprehension, and error handling in tools.

Read Article →

Part 7

MCP Servers & Built-In Tools

40 min read

MCPTransportCCA Domain 2

Model Context Protocol architecture, transport layers (stdio, SSE), dynamic tool discovery, and built-in tools.

Read Article →

Part 8

CLAUDE.md, Rules & Skills

35 min read

CLAUDE.mdRulesCCA Domain 3

Instruction hierarchy (CLAUDE.md), behavioral rules, skill definitions, and allowedTools restrictions.

Read Article →

Part 9

Plan Mode, Refinement & CI/CD

35 min read

Plan ModeCI/CDHeadless

Plan mode for design-before-code, iterative refinement loops, and headless CI/CD integration.

Read Article →

Part 10

Prompt Precision & Few-Shot

35 min read

PromptsFew-ShotCCA Domain 4

System prompt engineering, few-shot examples, negative constraints, and output steering techniques.

Read Article →

Part 11

Structured Output & Batch

35 min read

JSON ModePydanticBatches API

Guaranteed JSON via tool_use trick, Pydantic validation, Message Batches API for bulk processing.

Read Article →

Part 12

Context Preservation

35 min read

200K TokensSummarizationMemory

Token management, summarization strategies, sliding windows, and memory architectures for long agents.

Read Article →

Part 13

Escalation, Errors & Human Review

35 min read

Error RecoveryHITLCCA Domain 5

Error recovery patterns, human-in-the-loop escalation, confidence routing, and approval workflows.

Read Article →

Part 14

Extended Thinking

30 min read

ThinkingBudget TokensCoT

Extended thinking for complex reasoning: budget tokens, streaming thinking blocks, and chain-of-thought.

Read Article →

Part 15

Agent Environments & Cloud

30 min read

SandboxDockerSecurity

Container sandboxing, cloud deployments, network isolation, and security boundaries for agents.

Read Article →

Part 16

Administration & Compliance

30 min read

API KeysCost ControlGovernance

API key management, usage monitoring, rate limits, cost controls, and organizational governance.

Read Article →

Part 17

Use Cases & Ticket Routing

18 min read

Routingtool_choiceClassification

Choosing the right Claude pattern, intent classification, tool_choice modes, and confidence-based routing.

Read Article →

Part 18

Customer Support Agent

22 min read

SupportEscalationVerification

End-to-end support agent with MCP tools, identity verification, escalation criteria, and human handoff.

Read Article →

Part 19

Moderation & Legal Summarization

20 min read

ModerationLegalCascading

Content moderation classifiers, human review routing, legal document chunking, and model cascading.

Read Article →

Part 20

Guardrails: Hallucinations & Consistency

19 min read

GroundingCitationsConsistency

Reduce hallucinations with source grounding, nullable fields, temperature control, and eval-driven consistency.

Read Article →

Part 21

Guardrails: Jailbreaks & Prompt Leak

18 min read

SecurityInjectionRed Team

Defense-in-depth against jailbreaks, input sanitization, prompt leak prevention, and red-team testing.

Read Article →

Part 22

Built-In Tools & Permissions

45 min read

Built-in ToolsPermissionscanUseTool

All 15 built-in tools, permission modes (acceptEdits, bypassPermissions, plan), evaluation order, scoped Bash rules, and canUseTool callback.

Read Article →

Part 23

Streaming, Structured Output & Deployment

50 min read

Streamingresult_schemaDocker

Real-time streaming UIs, structured output via result_schema, file checkpointing, cost tracking, OpenTelemetry, and Docker/serverless deployment.

Read Article →

14-Part Google Gemini SDK Track

Google Gemini SDK & Interactions API

Master the modern Google GenAI ecosystem — from generateContent fundamentals through the Interactions API, Antigravity managed agents, Deep Research, Computer Use, Live API streaming, and enterprise production patterns. Covers Gemini 3.5 Flash, multimodal media, and the full tool suite.

Part 1

Platform Setup & SDK Fundamentals

35 min read

google-genaiAPI KeysModel Directory

Install the google-genai SDK, configure API keys, explore the model directory (3.5 Flash, 3.1 Pro, Nano Banana), pricing tiers, and MCP coding agent setup.

Read Article →

Part 2

Text Generation & Structured Outputs

35 min read

generateContentJSON SchemaStreaming

Generate text with generateContent, enforce type-safe JSON via response_format, streaming responses, token counting, and long context handling.

Read Article →

Part 3

Multimodal — Vision, Video & Documents

40 min read

VisionVideoFiles API

Image understanding vs. generation, video processing with timestamps, document/PDF analysis, media resolution settings, and the Files API for large uploads.

Read Article →

Part 4

Speech, Audio & Generative Media

35 min read

TTSVeoImagenLyria

Speech generation, audio understanding, expressive TTS tags, Veo video synthesis, Imagen image generation, and Lyria 3 music creation.

Read Article →

Part 5

Thinking, Reasoning & Thought Signatures

35 min read

thinkingBudgetSignaturesDeep Reasoning

Deep reasoning as default behavior, thinking_budget control, encrypted thought signatures, multi-turn preservation, and cost implications.

Read Article →

Part 6

Function Calling & Tool Integration

40 min read

Function CallingType HintsParallel Tools

Tool declarations, strict response matching, multimodal function responses, parallel tool calls, and combining tools with function calling.

Read Article →

Part 7

Built-in Tools: Search, Maps & Code

35 min read

Google SearchMapsURL Context

Google Search grounding, Maps grounding, URL context tool, code execution sandbox, combining multiple built-in tools, and rendering grounding metadata.

Read Article →

Part 8

File Search & RAG

35 min read

FileSearchStoresEmbedding 2RAG

FileSearchStores, Gemini Embedding 2, document management, semantic retrieval, and combining File Search with generation.

Read Article →

Part 9

The Interactions API (Beta)

45 min read

InteractionsStepsStateful

The architectural shift from generateContent to stateful Interactions, the steps timeline, previous_interaction_id, polymorphic response_format, and migration guide.

Read Article →

Part 10

Autonomous Agents & Antigravity SDK

45 min read

AntigravityManaged AgentsHooks

Antigravity Agent managed sandboxes, custom agents with inline environments, hook interception engine, system instructions, and multimodal agent inputs.

Read Article →

Part 11

Deep Research & Computer Use

40 min read

Deep ResearchComputer UseBrowser

Deep Research Agent with background mode and citations, Computer Use browser automation loops, observe-think-act cycle, and safety policy evaluation.

Read Article →

Part 12

Live API & Real-Time Streaming

40 min read

WebSocketEphemeral TokensBidirectional

Bidirectional WebSocket architecture, ai.live.connect(), session management, ephemeral tokens for client-side security, and real-time audio processing.

Read Article →

Part 13

Optimization & Production Operations

35 min read

CachingFlex InferenceBatch API

Context caching (explicit + implicit), Flex and Priority inference, Batch API, Webhooks, data logging, PII management, and billing optimization.

Read Article →

Part 14

Enterprise Migration & Frameworks

35 min read

MigrationGCP OAuthLangChain

May 2026 migration checklist, GCP OAuth service accounts, IAM, LangChain/LangGraph, CrewAI, LlamaIndex, Vercel AI SDK, and OpenAI compatibility.

Read Article →

14-Part PydanticAI SDK Track

PydanticAI SDK & Ecosystem

Master the PydanticAI framework — type-safe agents with dependency injection, multi-provider model support, function tools & toolsets, Pydantic Graph for workflows, Pydantic Evals for evaluation, MCP integration, durable execution with Temporal/DBOS/Prefect, and production-ready patterns with Logfire observability.

Part 1

Platform Setup & First Agent

35 min read

InstallationFirst AgentGateway

Install PydanticAI, configure the Gateway, create your first type-safe agent, understand the agent lifecycle, and explore troubleshooting patterns.

Read Article →

Part 2

Agents, Dependencies & Output

40 min read

AgentsDIOutput Types

Agent architecture, dependency injection with typed Deps, structured output types with Pydantic models, and capabilities configuration.

Read Article →

Part 3

Messages, Chat History & Model Requests

35 min read

MessagesChat HistoryDirect Requests

Message types and chat history management, direct model requests for low-level control, conversation patterns, and stateful interactions.

Read Article →

Part 4

Models & Multi-Provider Support

40 min read

14 ProvidersOpenAIAnthropicOllama

Configure all 14 providers — OpenAI, Anthropic, Google, xAI, Bedrock, Cerebras, Cohere, Groq, HuggingFace, Mistral, Ollama, OpenRouter, and Outlines.

Read Article →

Part 5

Function Tools & Toolsets

40 min read

Function ToolsToolsetsDeferred

Define function tools with type annotations, advanced tool features (retries, prepare), toolsets for grouped capabilities, and deferred tool loading.

Read Article →

Part 6

Native, Common & Third-Party Tools

35 min read

Native ToolsCommon ToolsThird-Party

Provider-native tools (code interpreter, search), common tool patterns (file I/O, HTTP), and integrating third-party tool libraries.

Read Article →

Part 7

Hooks, Agent Specs & Extensibility

35 min read

HooksAgent SpecsExtensibility

Hook lifecycle (pre/post tool call, model request), agent specs for documentation and testing, and extensibility patterns for custom behaviors.

Read Article →

Part 8

Multimodal Input & Thinking

35 min read

ImagesAudioVideoThinking

Image, audio, video, and document inputs in agents. Thinking/reasoning mode configuration, HTTP request retry strategies, and error handling.

Read Article →

Part 9

Multi-Agent Patterns & Testing

40 min read

Multi-AgentTestingWeb Chat

Multi-agent orchestration patterns, agent delegation, Web Chat UI integration, embeddings, and comprehensive testing strategies for AI agents.

Read Article →

Part 10

MCP Integration

40 min read

MCP ClientFastMCPMCP Server

Model Context Protocol integration — MCP client setup, FastMCP client for rapid development, building MCP servers, and tool interoperability.

Read Article →

Part 11

Pydantic Graph

40 min read

Graph BuilderStepsParallel

Graph builder for complex workflows — defining steps, joins & reducers, decision nodes, parallel execution, and composing graphs with agents.

Read Article →

Part 12

Pydantic Evals

45 min read

EvaluatorsLLM JudgeOnline Eval

Evaluation framework — built-in evaluators, LLM judge, custom evaluators, report & span-based evaluation, online eval, dataset management, and concurrency.

Read Article →

Part 13

Integrations & UI Streams

40 min read

LogfireTemporalAG-UIA2A

Logfire observability, durable execution (Temporal, DBOS, Prefect, Restate), UI event streams (AG-UI, Vercel AI), Agent2Agent protocol, and CLI.

Read Article →

Part 14

Harness, Gateway & Production

40 min read

HarnessCode ModeGateway

PydanticAI Harness and Code Mode for coding agents, Gateway configuration, Web Chat UI, embeddings, coding agent skills, and production deployment patterns.

Read Article →

14-Part CrewAI SDK Track

CrewAI SDK & Multi-Agent Orchestration

Master CrewAI’s production-grade multi-agent framework — agents with roles and goals, tasks and processes, Flows for stateful orchestration, 75+ built-in tools, MCP integration, knowledge and memory systems, human-in-the-loop workflows, observability with 15+ platforms, and migration patterns from LangGraph.

Part 1

Platform Setup & First Crew

35 min read

InstallationCLIQuickstart

Install CrewAI, scaffold projects with the CLI, build your first crew with Build with AI, and understand the project structure and execution model.

Read Article →

Part 2

Agents & Capabilities

40 min read

AgentsLLMsSkills

Define agents with roles, goals, and backstories. Configure LLMs, agent capabilities (tools, MCP, skills, knowledge), and craft effective agent personas.

Read Article →

Part 3

Tasks & Processes

40 min read

TasksSequentialHierarchical

Define tasks with descriptions and expected outputs. Orchestrate with sequential and hierarchical processes, conditional tasks, and agent collaboration.

Read Article →

Part 4

Crews & Production Architecture

40 min read

CrewsProductionAsync

Build production-grade crews, async kickoff patterns, kickoff-for-each, custom manager agents, and the recommended production architecture.

Read Article →

Part 5

Flows & State Management

45 min read

Flows@start@listenState

Build event-driven workflows with @start, @listen, and @router decorators. Master structured and unstructured state, persistence with LanceDB, and state restoration.

Read Article →

Part 6

Tools & Custom Tools

40 min read

BaseToolCustomPublish

Use built-in tools, create custom tools with BaseTool, publish to the tool registry, coding tools, and force tool output as result patterns.

Read Article →

Part 7

Tool Categories & Ecosystem

35 min read

75+ ToolsWebDatabase

Explore 75+ built-in tools across 8 categories: File/Document, Web Scraping, Search/Research, Database, AI/ML, Cloud/Storage, Integrations, and Automation.

Read Article →

Part 8

MCP Integration

40 min read

MCP DSLStdioSSEHTTP

Integrate MCP servers with the simple DSL syntax, Stdio/SSE/HTTP transports, connect to multiple servers, and security best practices.

Read Article →

Part 9

Knowledge, Memory & Files

40 min read

KnowledgeMemoryTraining

Add knowledge from PDFs, CSVs, and URLs. Configure memory systems (short-term, long-term, entity), file handling, and agent training.

Read Article →

Part 10

Planning, Reasoning & Testing

35 min read

PlanningReasoningTesting

Enable agent planning and reasoning capabilities, test crews systematically, replay tasks from latest kickoff, and validate agent behavior.

Read Article →

Part 11

Human-in-the-Loop & Multimodal

35 min read

HITLMultimodalDALL-E

Implement human-in-the-loop workflows, human input on execution, feedback in Flows, multimodal agents, and image generation with DALL-E.

Read Article →

Part 12

Advanced Patterns & Hooks

40 min read

HooksEventsCheckpoints

Customize prompts, fingerprinting, event listeners, checkpointing, execution hooks (LLM call hooks, tool call hooks), and annotations in crew.py.

Read Article →

Part 13

Observability & Telemetry

40 min read

TracingLangfuseDatadog

Built-in tracing with CrewAI AMP, plus integrations with Langfuse, Datadog, Arize Phoenix, Opik, MLflow, Portkey, Weave, and 10+ more platforms.

Read Article →

Part 14

Migration, Evaluation & Production

40 min read

LangGraphMigrationLLM Selection

Migrate from LangGraph to CrewAI, evaluate use cases, strategic LLM selection, upgrading versions, and using CrewAI without LiteLLM.

Read Article →

Cookie Consent

Foundation Track — SDK-Agnostic

LangChain & LangGraph SDK Deep Dive

OpenAI SDK Deep Dive

Anthropic Claude SDK & CCA Certification

Google Gemini SDK & Interactions API

PydanticAI SDK & Ecosystem

CrewAI SDK & Multi-Agent Orchestration