Memory systems, multi-agent orchestration, RAG pipelines, and production-grade agent infrastructure — the advanced architecture skills for engineers who are past the basics.
What Is MEMORY.md? Long-Term Brain for AI Agents
How agents maintain state and context across sessions.
Karpathy LLM Wiki: The Pattern Behind Agent Memory
Andrej Karpathy's approach to building persistent agent memory.
RAG vs Agentic RAG: Why Search Beats Embeddings for Code
When to move beyond naive RAG to agentic retrieval.
Langflow: Build Visual RAG Pipelines and Multi-Agent Workflows
Visual orchestration of complex agent pipelines.
Headroom: Context Compression for AI Agents
Keep agents effective even when context windows fill up.
Prompt Caching: LLM Cost, Latency, and Security Framework
Cache prompts intelligently to cut costs without sacrificing freshness.
Self-Harness: AI Agents That Improve Their Own Framework
The research pushing toward self-improving agent scaffolding.
Search as Code: Rethinking Search for the Agentic Era
How agentic search differs from keyword retrieval.
CocoIndex: Incremental Indexing for Always-Fresh Agent Context
Keep agent knowledge bases in sync without full reindexing.
Multi-Agent Orchestration Patterns
Orchestrator/worker, pipelines, fan-out, debate — the five patterns for production agent systems.
From AGI to ASI: DeepMind's 4 Pathways
The 57-page roadmap for what comes after human-level AI.
This pathway assumes you already understand AI agent basics (what loops, tools, and harnesses are) and are ready to tackle production concerns: memory systems that persist across sessions, agentic RAG pipelines for dynamic knowledge retrieval, context compression for long-running agents, multi-agent orchestration patterns, and prompt caching for cost optimization at scale.
Multi-agent orchestration is the design of systems where multiple AI agents collaborate to complete tasks too large or complex for a single agent. Common patterns include orchestrator/worker (one agent coordinates many), pipelines (agents pass work sequentially), fan-out (parallel specialist agents), and debate (agents challenge each other's outputs). This pathway covers all major patterns with production implementation guidance.
11 articles, approximately 8 hours. This is the deepest technical pathway on the platform and is recommended after completing Building AI Agents.
Understand what AI actually is — tokens, transformers, agents, and the landscape. Start here if you're new.
10 articles · ~4h →Go from vague requests to precise, reproducible AI outputs. The skill that underpins everything.
11 articles · ~5h →Go from zero to productive with Claude Code — the terminal AI coding agent that ships real projects.
13 articles · ~7h →Understand and build the loops, harnesses, and protocols that make AI agents reliable and autonomous.
11 articles · ~6h →Practical AI adoption for your specific function — marketing, engineering, HR, finance, and more.
10 articles · ~4h →Navigate the crowded model market — Claude, GPT, Gemini, open-source — and understand the tradeoffs.
10 articles · ~6h →
