AIML System Design
s09g
Ex-Meta, Ex-Google TL
This course is popular
45 people enrolled last week.
AI 系统设计与架构决策
Language of instruction: Mandarin Chinese
面向工程师、技术型 PM 和 AI Builder 的 AI 系统设计课程。用真实 AI 产品案例，为你建立一套可迁移的架构心智模型：如何澄清需求、定义 LLM 目标、划分系统边界、设计 RAG / Agent / 多模态 / 推理架构，并在延迟、成本、质量、安全和可观测性之间做取舍。
Design AI Systems. Explain Your Decisions.
Learn AI system design through document Q&A, customer support agents, multimodal assistants, and AI coding tools.
Each case starts with product requirements and a simple architecture. We then work through the decisions that shape the system: how to provide context, choose between workflows and agents, estimate capacity and cost, evaluate quality, and recover from failures.
Build the engineering judgment to choose a design, explain its trade-offs, and recognize when it needs to evolve. These skills carry over to both engineering work and AI system design interviews.
What you’ll learn
学会用一套可复用框架，拆解真实 GenAI 产品从需求、架构到评测和上线的关键决策。 Build the judgment to choose, explain, and evolve an AI architecture.
学会从用户体验、业务目标、系统边界、模型能力、数据流、评测、成本、延迟和安全等维度拆解 AI 产品问题
Turn requirements into an architecture, compare alternatives, and explain the evidence behind each decision.
判断什么时候该用 workflow，什么时候该用 agent，以及如何设计 tool calling、memory、state management、human-in-the-loop、fallback 和 escalation
Choose the right level of autonomy; design tool use, state, human review, and recovery together.
学会定义 AI 产品的成功指标，构建 eval dataset，结合 human rating、LLM judge、online metrics 和用户反馈持续迭代系统
Define what good looks like, identify the failing stage, and test whether a change improves the product.
Learn directly from s09g
s09g
Staff engineer, PyTorch/ExecuTorch Committer, Google Kubernetes Contributor
Who this course is for
正在准备 AI / LLM 系统设计面试，想把 RAG、Agent、Eval、多模态讲清楚的后端 / 全栈工程师
Engineers building AI systems or preparing for design interviews.
正在做 AI 产品或创业项目，已有 demo / prototype，但不知道如何推进到 production 的 AI Builder
Builders making architecture choices beyond the prototype.
正在公司内部推动 AI 项目，需要评审架构方案、控制成本质量风险的 Tech Lead / Technical PM
Leads reviewing AI architecture, cost, and quality.
Prerequisites
软件开发经验｜Software Development
有后端、全栈或产品技术背景，理解 API、数据库、异步任务和基础系统设计概念。 Know APIs, databases, async tasks, and basic system design.
系统设计知识｜System Design Basics
不要求机器学习研究背景，但需要愿意从目标、数据、评测和工程取舍角度分析 AI 系统。 No ML research background required; use an engineering lens.
LLM 使用经验｜LLM Experience
适合已经接触过 ChatGPT / Claude / Cursor，并希望系统掌握 AI 架构方法的人。 Have used ChatGPT, Claude, or Cursor.
What's included
Live sessions
Learn directly from s09g in a real-time, interactive format.
自主学习｜Self-paced Learning
这是一门 self-paced 课程。你可以按自己的节奏学习录播内容、查看案例材料和完成架构练习。 Recorded lessons, case materials, and architecture exercises you can study at your own pace.
课程回放与资料｜Recordings & Materials
可反复查看课程录像、课件、架构图和案例拆解。 Revisit the explanations, diagrams, and design reasoning as you apply them to new problems.
AI 系统案例｜AI System Case Studies
围绕 Chat with Docs、客服 Agent、代码助手、多模态推理和 NotebookLM 等案例拆解。 Trace design decisions across knowledge assistants, support agents, coding tools, and multimodal systems.
Maven Guarantee
Your purchase is backed by the Maven Guarantee.
Course syllabus
Live sessions
1 hr / week
Projects
1 hr / week
Async content
2 hrs / week
Week 1
LLM 推理基础｜LLM Inference
文档问答与 RAG｜Chat with Docs
客服 Agent｜Customer Support Agents
Week 2
多模态推理｜Multimodal Inference
视频生成｜Text-to-Video Generation
Frequently asked questions
Maven for Teams
Reimbursement
Get your company to pay
Everything L&D needs: email template, receipts, and certificate of completion.
Get reimbursedTeam discount
Learn with your teammates
Save 20%+ when 2 or more teammates enroll in the same cohort.
Save 20%+ with a teamPrivate cohort
Run a cohort for your org
A dedicated cohort with a custom schedule and curriculum, tailored to your team.
Book a private cohort$699
USD