Learn Advanced Agentic AI Engineering & Autonomous Software Engineering
// Skill map, learning contract, target performance, and deliberate practice plan for mastering advanced agentic AI engineering and autonomous software engineering.
This overview is designed to help you choose the right entry point quickly. Follow the full track from lesson one, continue from your last checkpoint, or jump straight into a phase that matches what you need right now.
Curriculum Map
Navigate by phase, then choose the lesson that matches your current depth.
Kaufman Skill Map for Agentic AI Engineering
24 minSkill map, learning contract, target performance, and deliberate practice plan for mastering advanced agentic AI engineering and autonomous software engineering.
Agentic AI First Principles
22 minFirst principles of agentic AI: definitions, taxonomy, autonomy, workflows versus agents, execution loops, and production decision criteria.
Autonomy Boundaries and Control Theory
16 minAutonomy boundaries, control surfaces, decision rights, blast radius, approval gates, and safe delegation models for production-grade agentic systems.
Agent Runtime Architecture
14 minRuntime architecture for production agentic systems: model adapter, planner, executor, tool gateway, policy engine, memory, state, persistence, evaluator, human-in-the-loop, and observability.
Agentic Workflow vs Agent Loop
18 minDistinguish deterministic workflows, LLM workflows, semi-agentic workflows, and autonomous agent loops, then learn when to choose each architecture in production systems.
Task Decomposition and Planning
15 minLearn how agentic systems decompose goals into bounded tasks, plans, dependencies, evidence, constraints, verification steps, and recovery paths.
Tool Calling Engineering
19 minEngineering-grade treatment of tool calling as capability contracts, including schema design, side-effect modeling, idempotency, policy gates, error semantics, observability, and safe execution.
MCP and Agent Integration Layer
18 minProduction-oriented guide to MCP and agent integration architecture, covering hosts, clients, servers, tools, resources, prompts, authorization, trust boundaries, gateway patterns, observability, and enterprise rollout.
Context Engineering
19 minContext engineering for production-grade agentic systems: context contracts, prompt layers, retrieval budgets, source prioritization, compression, stale-context control, and context observability.
Memory Architecture
17 minMemory architecture for agentic systems: working, episodic, semantic, procedural, and long-term memory; memory lifecycle, retrieval, poisoning defenses, retention, and auditability.
RAG for Agentic Systems
21 minAgentic RAG as an evidence, retrieval, grounding, and verification layer for production-grade agents and autonomous software engineering systems.
Agent State Machines
17 minExplicit state-machine design for reliable, replayable, pauseable, auditable, and policy-governed agentic systems.
Human-in-the-Loop and Approval Gates
22 minHuman-in-the-loop, approval gates, escalation, review policy, and decision-rights architecture for safe agentic systems.
Multi-Agent Systems
20 minMulti-agent system design, role specialization, coordination topology, shared state, communication contracts, failure modes, and production evaluation.
Agent Communication Protocols
19 minAgent communication protocols, message envelopes, task handoff, provenance, structured outputs, trust boundaries, and versioning for production agentic systems.
Agentic Design Patterns
21 minProduction-ready agentic design patterns: router, planner-executor, evaluator-optimizer, verifier, tool firewall, policy proxy, sandboxed actor, memory curator, and autonomous SWE patterns.
Agentic Anti-Patterns
26 minFailure catalog untuk agentic systems: over-agentification, fake autonomy, prompt spaghetti, unbounded loops, blind tool trust, hidden state, memory abuse, unverifiable completion, dan cara mendeteksi serta meremediasinya.
Autonomous Software Engineering Foundations
19 minFoundations of autonomous software engineering: issue intake, repository understanding, reproduction, patch planning, code editing loop, test verification, PR evidence packet, review gates, and production-grade coding agent lifecycle.
Repository Understanding Agents
19 minRepository understanding agents for autonomous software engineering: repo mapping, symbol search, build graph discovery, test graph discovery, dependency analysis, ownership, conventions, risk surface, and context packet generation.