AI Engineering from Scratch（英文原版）
You do not need to scan 523 lessons before beginning. Pick one goal. Each link opens the same curriculum on GitHub or the website, and both versions use the same lesson code.
课时
- Setup & Tooling
- Math Foundations
- Linear Algebra Intuition
- Vectors, Matrices & Operations
- Matrix Transformations & Eigenvalues
- Calculus for ML: Derivatives & Gradients
- Chain Rule & Automatic Differentiation
- Probability & Distributions
- Bayes' Theorem & Statistical Thinking
- Optimization: Gradient Descent Family
- Information Theory: Entropy, KL Divergence
- Dimensionality Reduction: PCA, t-SNE, UMAP
- Singular Value Decomposition
- Tensor Operations
- Numerical Stability
- Norms & Distances
- Statistics for ML
- Sampling Methods
- Linear Systems
- Complex Numbers for AI
- The Fourier Transform
- Graph Theory for ML
- Stochastic Processes
- ML Fundamentals
- What Is Machine Learning
- Linear Regression from Scratch
- Logistic Regression & Classification
- Decision Trees & Random Forests
- Support Vector Machines
- KNN & Distance Metrics
- Unsupervised Learning: K-Means, DBSCAN
- Feature Engineering & Selection
- Model Evaluation: Metrics, Cross-Validation
- Bias, Variance & the Learning Curve
- Ensemble Methods: Boosting, Bagging, Stacking
- Hyperparameter Tuning
- ML Pipelines & Experiment Tracking
- Naive Bayes
- Time Series Fundamentals
- Anomaly Detection
- Handling Imbalanced Data
- Feature Selection
- Deep Learning Core
- The Perceptron: Where It All Started
- Multi-Layer Networks & Forward Pass
- Backpropagation from Scratch
- Activation Functions: ReLU, Sigmoid, GELU & Why
- Loss Functions: MSE, Cross-Entropy, Contrastive
- Optimizers: SGD, Momentum, Adam, AdamW
- Regularization: Dropout, Weight Decay, BatchNorm
- Weight Initialization & Training Stability
- Learning Rate Schedules & Warmup
- Build Your Own Mini Framework
- Introduction to PyTorch
- Introduction to JAX
- Debugging Neural Networks
- Computer Vision
- Image Fundamentals: Pixels, Channels, Color Spaces
- Convolutions from Scratch
- CNNs: LeNet to ResNet
- Image Classification
- Transfer Learning & Fine-Tuning
- Object Detection — YOLO from Scratch
- Semantic Segmentation — U-Net
- Instance Segmentation — Mask R-CNN
- Image Generation — GANs
- Image Generation — Diffusion Models
- Stable Diffusion — Architecture & Fine-Tuning
- Video Understanding — Temporal Modeling
- 3D Vision: Point Clouds, NeRFs
- Vision Transformers (ViT)
- Real-Time Vision: Edge Deployment
- Build a Complete Vision Pipeline
- Self-Supervised Vision — SimCLR, DINO, MAE
- Open-Vocabulary Vision — CLIP
- OCR & Document Understanding
- Image Retrieval & Metric Learning
- Keypoint Detection & Pose Estimation
- 3D Gaussian Splatting from Scratch
- Diffusion Transformers & Rectified Flow
- SAM 3 & Open-Vocabulary Segmentation
- Vision-Language Models (ViT-MLP-LLM)
- Monocular Depth & Geometry Estimation
- Multi-Object Tracking & Video Memory
- World Models & Video Diffusion
- NLP: Foundations to Advanced
- Text Processing: Tokenization, Stemming, Lemmatization
- Bag of Words, TF-IDF & Text Representation
- Word Embeddings: Word2Vec from Scratch
- GloVe, FastText & Subword Embeddings
- Sentiment Analysis
- Named Entity Recognition (NER)
- POS Tagging & Syntactic Parsing
- Text Classification — CNNs & RNNs for Text
- Sequence-to-Sequence Models
- Attention Mechanism — The Breakthrough
- Machine Translation
- Text Summarization
- Question Answering Systems
- Information Retrieval & Search
- Topic Modeling: LDA, BERTopic
- Text Generation
- Chatbots: Rule-Based to Neural
- Multilingual NLP
- Subword Tokenization: BPE, WordPiece, Unigram, SentencePiece
- Structured Outputs & Constrained Decoding
- NLI & Textual Entailment
- Embedding Models Deep Dive
- Chunking Strategies for RAG
- Coreference Resolution
- Entity Linking & Disambiguation
- Relation Extraction & Knowledge Graph Construction
- LLM Evaluation: RAGAS, DeepEval, G-Eval
- Long-Context Evaluation: NIAH, RULER, LongBench, MRCR
- Dialogue State Tracking
- Speech & Audio
- Audio Fundamentals: Waveforms, Sampling, FFT
- Spectrograms, Mel Scale & Audio Features
- Audio Classification
- Speech Recognition (ASR)
- Whisper: Architecture & Fine-Tuning
- Speaker Recognition & Verification
- Text-to-Speech (TTS)
- Voice Cloning & Voice Conversion
- Music Generation
- Audio-Language Models
- Real-Time Audio Processing
- Build a Voice Assistant Pipeline
- Neural Audio Codecs — EnCodec, SNAC, Mimi, DAC
- Voice Activity Detection & Turn-Taking
- Streaming Speech-to-Speech — Moshi, Hibiki
- Voice Anti-Spoofing & Audio Watermarking
- Audio Evaluation — WER, MOS, MMAU, Leaderboards
- Transformers Deep Dive
- Why Transformers: The Problems with RNNs
- Self-Attention from Scratch
- Multi-Head Attention
- Positional Encoding: Sinusoidal, RoPE, ALiBi
- The Full Transformer: Encoder + Decoder
- BERT — Masked Language Modeling
- GPT — Causal Language Modeling
- T5, BART — Encoder-Decoder Models
- Vision Transformers (ViT)
- Audio Transformers — Whisper Architecture
- Mixture of Experts (MoE)
- KV Cache, Flash Attention & Inference Optimization
- Scaling Laws
- Build a Transformer from Scratch
- Attention Variants — Sliding Window, Sparse, Differential
- Speculative Decoding — Draft, Verify, Repeat
- Generative AI
- Generative Models: Taxonomy & History
- Autoencoders & VAE
- GANs: Generator vs Discriminator
- Conditional GANs & Pix2Pix
- StyleGAN
- Diffusion Models — DDPM from Scratch
- Latent Diffusion & Stable Diffusion
- ControlNet, LoRA & Conditioning
- Inpainting, Outpainting & Editing
- Video Generation
- Audio Generation
- 3D Generation
- Flow Matching & Rectified Flows
- Evaluation: FID, CLIP Score
- Visual Autoregressive Modeling (VAR): Next-Scale Prediction
- Reinforcement Learning
- LLMs from Scratch
- Tokenizers: BPE, WordPiece, SentencePiece
- Building a Tokenizer from Scratch
- Data Pipelines for Pre-Training
- Pre-Training a Mini GPT (124M)
- Distributed Training, FSDP, DeepSpeed
- Instruction Tuning — SFT
- RLHF — Reward Model + PPO
- DPO — Direct Preference Optimization
- Constitutional AI & Self-Improvement
- Evaluation — Benchmarks, Evals
- Quantization: INT8, GPTQ, AWQ, GGUF
- Inference Optimization
- Building a Complete LLM Pipeline
- Open Models: Architecture Walkthroughs
- Speculative Decoding and EAGLE-3
- Differential Attention (V2)
- Native Sparse Attention (DeepSeek NSA)
- Multi-Token Prediction (MTP)
- DualPipe Parallelism
- DeepSeek-V3 Architecture Walkthrough
- Jamba — Hybrid SSM-Transformer
- Async and Hogwild! Inference
- Speculative Decoding and EAGLE
- Gradient Checkpointing and Activation Recomputation
- LLM Engineering
- Prompt Engineering: Techniques & Patterns
- Few-Shot, CoT, Tree-of-Thought
- Structured Outputs
- Embeddings & Vector Representations
- Context Engineering
- RAG: Retrieval-Augmented Generation
- Advanced RAG: Chunking, Reranking
- Fine-Tuning with LoRA & QLoRA
- Function Calling & Tool Use
- Evaluation & Testing
- Caching, Rate Limiting & Cost
- Guardrails & Safety
- Building a Production LLM App
- Model Context Protocol (MCP)
- Prompt Caching & Context Caching
- Agent State Machines — Graphs, Nodes, Checkpoints
- Agent Framework Tradeoffs
- Multimodal AI
- Vision Transformers and the Patch-Token Primitive
- CLIP and Contrastive Vision-Language Pretraining
- BLIP-2 Q-Former as Modality Bridge
- Flamingo and Gated Cross-Attention
- LLaVA and Visual Instruction Tuning
- Any-Resolution Vision — Patch-n'-Pack and NaFlex
- Open-Weight VLM Recipes: What Actually Matters
- LLaVA-OneVision: Single, Multi, Video
- Qwen-VL Family and Dynamic-FPS Video
- InternVL3 Native Multimodal Pretraining
- Chameleon Early-Fusion Token-Only
- Emu3 Next-Token Prediction for Generation
- Transfusion Autoregressive + Diffusion
- Show-o Discrete-Diffusion Unified
- Janus-Pro Decoupled Encoders
- MIO Any-to-Any Streaming
- Video-Language Temporal Grounding
- Long-Video at Million-Token Context
- Audio-Language Models: Whisper to AF3
- Omni Models: Thinker-Talker Streaming
- Embodied VLAs: RT-2, OpenVLA, π0, GR00T
- Document and Diagram Understanding
- ColPali Vision-Native Document RAG
- Multimodal RAG and Cross-Modal Retrieval
- Multimodal Agents and Computer-Use (Capstone)
- Tools & Protocols
- The Tool Interface
- Function Calling Deep Dive
- Parallel and Streaming Tool Calls
- Structured Output
- Tool Schema Design
- MCP Fundamentals: Stateless Requests and JSON-RPC
- Building an MCP Server: Stateless Python and TypeScript
- Building an MCP Client: Discovery, Routing, and Dual-Era Fallback
- MCP Transports: stdio and Stateless Streamable HTTP
- MCP Resources and Prompts: Addressable Context for Stateless Servers
- MCP Model Input: Sampling Migration and Stateless MRTR
- Explicit Scope and Stateless Elicitation
- MCP Tasks Extension: Durable Work on a Stateless Core
- MCP Apps on the Stateless Protocol
- MCP Security: Poisoned Metadata, Routing, and MRTR State
- MCP Authorization: CIMD, Issuer Binding, PKCE, and Step-Up
- Stateless MCP Gateways and Registry Admission
- MCP Auth in Production: Issuer-Bound Enrollment and Tokens
- A2A Protocol
- OpenTelemetry GenAI
- LLM Routing Layer
- Agent Skills: Portable Contract and Runtime Boundary
- Capstone: Stateless Tool Ecosystem
- Skill Discovery and Progressive Disclosure
- Skill Invocation and Routing
- Skill Permissions, Sandboxes, and Trust
- Skill Evals, Packaging, and Portability
- MCP Tool Contracts and Content
- MCP Reliability, Cancellation, and Flow Control
- MCP Registry Supply Chain: Admission, Drift, and Rollback
- MCP Conformance Engineering: Versioning, Evidence, and Operations
- Agent Engineering
- The Agent Loop
- ReWOO and Plan-and-Execute
- Reflexion and Verbal Reinforcement Learning
- Tree of Thoughts and LATS
- Self-Refine and CRITIC
- Tool Use and Function Calling
- Agent Memory — Virtual Context and Memory Paging
- Memory Blocks and Sleep-Time Compute
- Hybrid Memory — Vector + Graph + KV
- Skill Libraries and Lifelong Learning (Voyager)
- Planning with HTN and Evolutionary Search
- Anthropic's Workflow Patterns
- Stateful Graph Orchestration — Durable Execution and Checkpoints
- The Actor Model for Agents
- Role-Based Agent Teams — Roles, Tasks, Processes
- OpenAI Agents SDK — Handoffs, Guardrails, Tracing
- The Harness as a Library — Subagents and Session Store
- Production Agent Runtimes
- Benchmarks — SWE-bench, GAIA, AgentBench
- Benchmarks — WebArena and OSWorld
- Computer Use — Claude, OpenAI CUA, Gemini
- Voice Agents — Pipecat and LiveKit
- OpenTelemetry GenAI Semantic Conventions
- Agent Observability — Langfuse, Phoenix, Opik
- Multi-Agent Debate and Collaboration
- Failure Modes — Why Agents Break
- Prompt Injection and the PVE Defense
- Orchestration Patterns — Supervisor, Swarm, Hierarchical
- Production Runtimes — Queue, Event, Cron
- Eval-Driven Agent Development
- Agent Workbench: Why Capable Models Still Fail
- The Minimal Agent Workbench
- Agent Instructions as Executable Constraints
- Repo Memory and Durable State
- Initialization Scripts for Agents
- Scope Contracts and Task Boundaries
- Runtime Feedback Loops
- Verification Gates
- Reviewer Agent: Separate Builder from Marker
- Multi-Session Handoff
- The Workbench on a Real Repo
- Capstone: Ship a Reusable Agent Workbench Pack
- Frame the Task Before the Agent Writes Code
- Build an Evidence-Backed Execution Plan
- Delegate Agent Work with Isolation and Merge Contracts
- Turn Every Agent Correction into a System Improvement
- Define the Outcome Before You Choose the Output
- Discover the Workflow People Actually Perform
- Map Assumptions and Resolve the Riskiest One First
- Choose the Smallest Slice That Can Change the Decision
- Write Specifications That Preserve Judgment
- Design Success Metrics Before the Result Exists
- Choose Prototype, Pilot, or Production Deliberately
- Build a Feedback Ratchet with Ownership and Retirement
- Autonomous Systems
- From Chatbots to Long-Horizon Agents (METR)
- STaR, V-STaR, Quiet-STaR: Self-Taught Reasoning
- AlphaEvolve: Evolutionary Coding Agents
- Darwin Gödel Machine: Self-Modifying Agents
- AI Scientist v2: Workshop-Level Research
- Automated Alignment Research (Anthropic AAR)
- Recursive Self-Improvement: Capability vs Alignment
- Bounded Self-Improvement Designs
- Autonomous Coding Agent Landscape (SWE-bench, CodeAct)
- Permission Modes for Autonomous Agents
- Browser Agents and Indirect Prompt Injection
- Durable Execution for Long-Running Agents
- Action Budgets, Iteration Caps, Cost Governors
- Kill Switches, Circuit Breakers, Canary Tokens
- HITL: Propose-Then-Commit
- Checkpoints and Rollback
- Constitutional AI and Rule Overrides
- Llama Guard and Input/Output Classification
- Anthropic Responsible Scaling Policy v3.0
- OpenAI Preparedness Framework and DeepMind FSF
- METR Time Horizons and External Evaluation
- CAIS, CAISI, and Societal-Scale Risk
- Multi-Agent & Swarms
- Why Multi-Agent
- FIPA-ACL Heritage and Speech Acts
- Communication Protocols
- The Multi-Agent Primitive Model
- Supervisor / Orchestrator-Worker Pattern
- Hierarchical Architecture and Decomposition Drift
- Society of Mind and Multi-Agent Debate
- Role Specialization — Planner / Critic / Executor / Verifier
- Parallel Swarm and Networked Architectures
- Group Chat and Speaker Selection
- Handoffs and Routines (Stateless Orchestration)
- A2A — The Agent-to-Agent Protocol
- Shared Memory and Blackboard Patterns
- Consensus and Byzantine Fault Tolerance
- Voting, Self-Consistency, and Debate Topology
- Negotiation and Bargaining
- Generative Agents and Emergent Simulation
- Theory of Mind and Emergent Coordination
- Swarm Optimization (PSO, ACO)
- MARL — MADDPG, QMIX, MAPPO
- Agent Economies, Token Incentives, Reputation
- Production Scaling — Queues, Checkpoints, Durability
- Failure Modes — MAST, Groupthink, Monoculture
- Evaluation and Coordination Benchmarks
- Case Studies and 2026 State of the Art
- Infrastructure & Production
- Managed LLM Platforms — Bedrock, Azure OpenAI, Vertex AI
- Inference Platform Economics — Fireworks, Together, Baseten, Modal
- GPU Autoscaling on Kubernetes — Karpenter, KAI Scheduler
- Serving Engine Internals — PagedAttention, Continuous Batching, Chunked Prefill
- EAGLE-3 Speculative Decoding in Production
- Prefix-Cache Serving — RadixAttention and KV Reuse
- Hardware-Specialized Inference Compilation — FP8 and NVFP4 on Blackwell
- Inference Metrics — TTFT, TPOT, ITL, Goodput, P99
- Production Quantization — AWQ, GPTQ, GGUF, FP8, NVFP4
- Cold Start Mitigation for Serverless LLMs
- Multi-Region LLM Serving and KV Cache Locality
- Edge Inference — ANE, Hexagon, WebGPU, Jetson
- LLM Observability Stack Selection
- Prompt Caching and Semantic Caching Economics
- Batch APIs — the 50% Discount as Industry Standard
- Model Routing as a Cost-Reduction Primitive
- Disaggregated Prefill/Decode — NVIDIA Dynamo and llm-d
- Production Serving Stack — KV Offloading and Cache-Aware Routing
- AI Gateways — LiteLLM, Portkey, Kong, Bifrost
- Shadow, Canary, and Progressive Deployment
- A/B Testing LLM Features — GrowthBook and Statsig
- Load Testing LLM APIs — k6, LLMPerf, GenAI-Perf
- SRE for AI — Multi-Agent Incident Response
- Chaos Engineering for LLM Production
- Security — Secrets, PII Scrubbing, Audit Logs
- Compliance — SOC 2, HIPAA, GDPR, EU AI Act, ISO 42001
- FinOps for LLMs — Unit Economics and Multi-Tenant Attribution
- Self-Hosted Serving Selection — Matching Engine to Hardware and Scale
- Ethics, Safety & Alignment
- Instruction-Following as Alignment Signal
- Reward Hacking & Goodhart's Law
- Direct Preference Optimization Family
- Sycophancy as RLHF Amplification
- Constitutional AI & RLAIF
- Mesa-Optimization & Deceptive Alignment
- Sleeper Agents — Persistent Deception
- In-Context Scheming in Frontier Models
- Alignment Faking
- AI Control — Safety Despite Subversion
- Scalable Oversight & Weak-to-Strong
- Red-Teaming: PAIR & Automated Attacks
- Many-Shot Jailbreaking
- ASCII Art & Visual Jailbreaks
- Indirect Prompt Injection
- Red-Team Tooling: Garak, Llama Guard, PyRIT
- WMDP & Dual-Use Capability Evaluation
- Frontier Safety Frameworks — RSP, PF, FSF
- Model Welfare Research
- Bias & Representational Harm
- Fairness Criteria: Group, Individual, Counterfactual
- Differential Privacy for LLMs
- Watermarking: SynthID, Stable Signature, C2PA
- Regulatory Frameworks: EU, US, UK, Korea
- EchoLeak & CVEs for AI
- Model, System & Dataset Cards
- Data Provenance & Training-Data Governance
- Alignment Research Ecosystem: MATS, Redwood, Apollo, METR
- Moderation Systems: OpenAI, Perspective, Llama Guard
- Dual-Use Risk: Cyber, Bio, Chem, Nuclear
- Capstone Projects
- Terminal-Native Coding Agent
- RAG over Codebase (Cross-Repo Semantic Search)
- Real-Time Voice Assistant (ASR → LLM → TTS)
- Multimodal Document QA (Vision-First)
- Autonomous Research Agent (AI-Scientist Class)
- DevOps Troubleshooting Agent for Kubernetes
- End-to-End Fine-Tuning Pipeline
- Production RAG Chatbot (Regulated Vertical)
- Code Migration Agent (Repo-Level Upgrade)
- Multi-Agent Software Engineering Team
- LLM Observability & Eval Dashboard
- Video Understanding Pipeline (Scene → QA)
- Stateless MCP Server with Registry and Governance
- Speculative-Decoding Inference Server
- Constitutional Safety Harness + Red-Team Range
- GitHub Issue-to-PR Autonomous Agent
- Personal AI Tutor (Adaptive, Multimodal)
- Agent Harness Loop Contract
- Tool Registry with Schema Validation
- JSON-RPC 2.0 Over Newline-Delimited Stdio
- Function Call Dispatcher
- Plan-Execute Control Flow
- Verification Gates and Observation Budget
- Sandbox Runner with Denylist and Path Jail
- Eval Harness with Fixture Tasks
- Observability with OTel GenAI Spans and Prometheus Metrics
- End-to-End Coding Agent on the Harness
- BPE Tokenizer From Scratch
- Tokenized Dataset with Sliding Window
- Token and Positional Embeddings
- Multi-Head Self-Attention
- Transformer Block from Scratch
- GPT Model Assembly
- Training Loop and Evaluation
- Loading Pretrained Weights
- Classifier Fine-Tuning by Head Swap
- Instruction Tuning by Supervised Fine-Tuning
- Direct Preference Optimization from Scratch
- Full Evaluation Pipeline
- Large Corpus Downloader
- HDF5 Tokenized Corpus
- Cosine LR with Linear Warmup
- Gradient Clipping and Mixed Precision
- Gradient Accumulation
- Checkpoint Save and Resume
- Distributed Data Parallel and FSDP from Scratch
- Language Model Evaluation Harness
- Hypothesis Generator
- Literature Retrieval
- Experiment Runner
- Result Evaluator
- Paper Writer
- Critic Loop
- Iteration Scheduler
- End-to-End Research Demo
- Vision Encoder Patches
- Vision Transformer Encoder
- Projection Layer for Modality Alignment
- Cross-Attention Fusion
- Vision-Language Pretraining
- Multimodal Evaluation
- Chunking Strategies, Compared
- Hybrid Retrieval with BM25 and Dense Embeddings
- Cross-Encoder Reranker
- Query Rewriting: HyDE, Multi-Query, and Decomposition
- RAG Evaluation: Precision, Recall, MRR, nDCG, Faithfulness, Answer Relevance
- End-to-End RAG System
- Task Spec Format
- Classical Metrics
- Code Exec Metric
- Perplexity and Calibration
- Leaderboard Aggregation
- End-to-End Eval Runner
- Collective Ops From Scratch
- Data Parallel DDP From Scratch
- ZeRO Optimizer State Sharding
- Pipeline Parallel and Bubble Analysis
- Sharded Checkpoint and Atomic Resume
- End-to-End Distributed Training
- Jailbreak Taxonomy
- Prompt Injection Detector
- Refusal Evaluation
- Content Classifier Integration
- Constitutional Rules Engine
- End-to-End Safety Gate
- Backers
- Forking Guide
- Lesson Template
- Roadmap
- Sponsorship
- .claude
- skills
- check-understanding
- claude-certification
- course-guide
- find-your-level
- Find Your Level
- references
- learn-agent-skills
- learn-mcp
- learn
- start-learning
- skills
- .github
- Book Pipeline
- Certification Curricula
- Claude Certification Curriculum
- Learn Claude Certifications From GitHub
- lessons
- 00-certification-strategy
- 01-claude-product-and-model-landscape
- 02-model-selection-and-token-economics
- 03-prompting-and-task-decomposition
- 04-context-knowledge-memory-and-caching
- 05-output-evaluation-and-validation
- 06-governance-safety-and-responsible-use
- 07-workflow-design-and-human-handoffs
- 08-messages-api-and-application-lifecycle
- 09-structured-output-and-defensive-parsing
- 10-tool-use-and-agentic-loops
- 11-mcp-server-design-and-integration
- 12-claude-agent-sdk-and-hooks
- 13-application-security-and-secrets
- 14-evals-testing-debugging-and-observability
- 15-claude-code-for-development-teams
- 16-multi-agent-orchestration-and-delegation
- 17-agent-sdk-sessions-subagents-and-context
- 18-tool-contracts-errors-and-progressive-discovery
- 19-claude-code-memory-rules-skills-and-ci
- 文档
- outputs
- Configuration Scope Audit: Python Service
- migration-review-skill
- 20-reliable-extraction-batch-and-reviewers
- 21-long-context-reliability-provenance-and-escalation
- 22-business-discovery-requirements-and-slas
- 23-end-to-end-architecture-and-value-tradeoffs
- 24-rag-retrieval-and-data-pipelines
- 25-integration-protocols-identity-and-least-privilege
- 26-production-observability-latency-and-cost
- 27-enterprise-governance-compliance-and-hitl
- 28-stakeholder-communication-adrs-and-lifecycle
- 29-associate-workflow-capstone
- 30-developer-application-capstone
- 31-architect-foundations-scenario-capstone
- 32-architect-professional-system-capstone
- references
- research
- Claude Certification Curriculum
- 文档
- Glossary
- AI Myths Busted
- AI Engineering Glossary
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- AI Engineering from Scratch（英文原版）
- i18n
- phases
- Phase 0: Setup & Tooling
- 01-dev-environment
- 04-apis-and-keys
- 05-jupyter-notebooks
- 09-data-management
- 12-debugging-and-profiling
- Phase 1: Math Foundations
- 01-linear-algebra-intuition
- 02-vectors-matrices-operations
- 03-matrix-transformations
- 04-calculus-for-ml
- outputs
- 05-chain-rule-and-autodiff
- 06-probability-and-distributions
- 07-bayes-theorem
- 08-optimization
- 09-information-theory
- outputs
- 10-dimensionality-reduction
- 11-singular-value-decomposition
- 12-tensor-operations
- 13-numerical-stability
- 14-norms-and-distances
- 15-statistics-for-ml
- outputs
- 16-sampling-methods
- outputs
- 17-linear-systems
- 18-convex-optimization
- outputs
- 19-complex-numbers
- 20-fourier-transform
- 21-graph-theory
- 22-stochastic-processes
- Phase 2: ML Fundamentals
- 01-what-is-machine-learning
- 02-linear-regression
- outputs
- 03-logistic-regression
- 04-decision-trees
- 05-support-vector-machines
- outputs
- 06-knn-and-distances
- 07-unsupervised-learning
- 08-feature-engineering
- outputs
- 09-model-evaluation
- outputs
- 10-bias-variance
- 11-ensemble-methods
- 12-hyperparameter-tuning
- 13-ml-pipelines
- 14-naive-bayes
- 15-time-series
- 16-anomaly-detection
- 17-imbalanced-data
- outputs
- 18-feature-selection
- outputs
- Phase 3: Deep Learning Core
- 01-the-perceptron
- outputs
- 02-multi-layer-networks
- 03-backpropagation
- 04-activation-functions
- 05-loss-functions
- 06-optimizers
- 07-regularization
- 08-weight-initialization
- 09-learning-rate-schedules
- 10-mini-framework
- 11-intro-to-pytorch
- 12-intro-to-jax
- 13-debugging-neural-networks
- 01-the-perceptron
- Phase 4: Computer Vision
- 01-image-fundamentals
- 02-convolutions-from-scratch
- 03-cnns-lenet-to-resnet
- 04-image-classification
- 05-transfer-learning
- 06-object-detection-yolo
- 07-semantic-segmentation-unet
- 08-instance-segmentation-mask-rcnn
- 09-image-generation-gans
- 10-image-generation-diffusion
- 11-stable-diffusion
- 12-video-understanding
- 13-3d-vision-nerf
- 14-vision-transformers
- 15-real-time-edge
- 16-vision-pipeline-capstone
- 17-self-supervised-vision
- 18-open-vocab-clip
- 19-ocr-document-understanding
- 20-image-retrieval-metric
- 21-keypoint-pose
- 22-3d-gaussian-splatting
- 23-diffusion-transformers-rectified-flow
- 24-sam3-open-vocab-segmentation
- 25-vision-language-models
- 26-monocular-depth
- 27-multi-object-tracking
- 28-world-models-video-diffusion
- Phase 5: NLP, Foundations to Advanced
- 01-text-processing
- 02-bag-of-words-tfidf
- 03-word-embeddings-word2vec
- 04-glove-fasttext-subword
- 05-sentiment-analysis
- 06-named-entity-recognition
- 07-pos-tagging-parsing
- 08-cnns-rnns-for-text
- 09-sequence-to-sequence
- 10-attention-mechanism
- 11-machine-translation
- 12-text-summarization
- 13-question-answering
- 14-information-retrieval-search
- 15-topic-modeling
- 16-text-generation-pre-transformer
- 17-chatbots-rule-to-neural
- 18-multilingual-nlp
- 19-subword-tokenization
- 20-structured-outputs-constrained-decoding
- 21-nli-textual-entailment
- 22-embedding-models-deep-dive
- 23-chunking-strategies-rag
- 24-coreference-resolution
- 25-entity-linking
- 26-relation-extraction-kg
- 27-llm-evaluation-frameworks
- 28-long-context-evaluation
- 29-dialogue-state-tracking
- Phase 6: Speech & Audio
- 01-audio-fundamentals
- 02-spectrograms-mel-features
- 03-audio-classification
- 04-speech-recognition-asr
- 05-whisper-architecture-finetuning
- 06-speaker-recognition-verification
- 07-text-to-speech
- 08-voice-cloning-conversion
- 09-music-generation
- 10-audio-language-models
- 11-real-time-audio-processing
- 12-voice-assistant-pipeline
- 13-neural-audio-codecs
- 14-voice-activity-detection-turn-taking
- 15-streaming-speech-to-speech-moshi-hibiki
- 16-anti-spoofing-audio-watermarking
- 17-audio-evaluation-metrics
- Phase 7: Transformers Deep Dive
- 01-why-transformers
- 02-self-attention-from-scratch
- 03-multi-head-attention
- 04-positional-encoding
- 05-full-transformer
- 06-bert-masked-language-modeling
- 07-gpt-causal-language-modeling
- 08-t5-bart-encoder-decoder
- 09-vision-transformers
- 10-audio-transformers-whisper
- 11-mixture-of-experts
- 12-kv-cache-flash-attention
- 13-scaling-laws
- 14-build-a-transformer-capstone
- 15-attention-variants
- outputs
- 16-speculative-decoding
- outputs
- Phase 8: Generative AI
- 01-generative-models-taxonomy-history
- 02-autoencoders-vae
- 03-gans-generator-discriminator
- 04-conditional-gans-pix2pix
- 05-stylegan
- 06-diffusion-ddpm-from-scratch
- 07-latent-diffusion-stable-diffusion
- 08-controlnet-lora-conditioning
- 09-inpainting-outpainting-editing
- 10-video-generation
- 11-audio-generation
- 12-3d-generation
- 13-flow-matching-rectified-flows
- 14-evaluation-fid-clip-score
- 19-visual-autoregressive-var
- Phase 9: Reinforcement Learning
- 01-mdps-states-actions-rewards
- 02-dynamic-programming
- 03-monte-carlo-methods
- 04-q-learning-sarsa
- 05-dqn
- 06-policy-gradients-reinforce
- 07-actor-critic-a2c-a3c
- 08-ppo
- 09-reward-modeling-rlhf
- 10-multi-agent-rl
- 11-sim-to-real-transfer
- 12-rl-for-games
- Phase 10: LLMs from Scratch
- 01-tokenizers
- 02-building-a-tokenizer
- outputs
- 03-data-pipelines
- 04-pre-training-mini-gpt
- outputs
- 05-scaling-distributed
- outputs
- 06-instruction-tuning-sft
- outputs
- 07-rlhf
- outputs
- 08-dpo
- outputs
- 09-constitutional-ai-self-improvement
- 10-evaluation
- 11-quantization
- 12-inference-optimization
- 13-building-complete-llm-pipeline
- 14-open-models-architecture-walkthroughs
- 15-speculative-decoding-eagle3
- 16-differential-attention-v2
- 17-native-sparse-attention
- 18-multi-token-prediction
- 19-dualpipe-parallelism
- 20-deepseek-v3-walkthrough
- 21-jamba-hybrid-ssm-transformer
- 22-async-hogwild-inference
- 25-speculative-decoding
- 34-gradient-checkpointing
- Phase 11: LLM Engineering
- 01-prompt-engineering
- 02-few-shot-cot
- 03-structured-outputs
- 04-embeddings
- 05-context-engineering
- 06-rag
- 07-advanced-rag
- 08-fine-tuning-lora
- 09-function-calling
- 10-evaluation
- 11-caching-cost
- 12-guardrails
- 13-production-app
- 14-model-context-protocol
- 15-prompt-caching
- 16-langgraph-state-machines
- 17-agent-framework-tradeoffs
- Phase 12: Multimodal AI
- 01-vision-transformer-patch-tokens
- 02-clip-contrastive-pretraining
- 03-blip2-qformer-bridge
- 04-flamingo-gated-cross-attention
- 05-llava-visual-instruction-tuning
- 06-any-resolution-patch-n-pack
- 07-open-weight-vlm-recipes
- 08-llava-onevision-single-multi-video
- 09-qwen-vl-family-dynamic-fps
- 10-internvl3-native-multimodal
- 11-chameleon-early-fusion-tokens
- 12-emu3-next-token-for-generation
- 13-transfusion-autoregressive-diffusion
- 14-show-o-discrete-diffusion-unified
- 15-janus-pro-decoupled-encoders
- 16-mio-any-to-any-streaming
- 17-video-language-temporal-grounding
- 18-long-video-million-token
- 19-audio-language-whisper-to-af3
- 20-omni-models-thinker-talker
- 21-embodied-vlas-openvla-pi0-groot
- 22-document-diagram-understanding
- 23-colpali-vision-native-rag
- 24-multimodal-rag-cross-modal
- 25-multimodal-agents-computer-use
- Phase 13: Tools & Protocols
- 01-the-tool-interface
- 02-function-calling-deep-dive
- 03-parallel-and-streaming-tool-calls
- 04-structured-output
- 05-tool-schema-design
- 06-mcp-fundamentals
- 07-building-an-mcp-server
- 08-building-an-mcp-client
- 09-mcp-transports
- 10-mcp-resources-and-prompts
- 11-mcp-sampling
- 12-mcp-roots-and-elicitation
- 13-mcp-async-tasks
- 14-mcp-apps
- 15-mcp-security-tool-poisoning
- 16-mcp-security-oauth-2-1
- 17-mcp-gateways-and-registries
- 18-mcp-auth-production
- 19-a2a-protocol
- 20-opentelemetry-genai
- 21-llm-routing-layer
- 22-skills-and-agent-sdks
- outputs
- skill-contract-reviewer
- outputs
- 23-capstone-tool-ecosystem
- 24-skill-discovery-and-progressive-disclosure
- outputs
- skill-catalog-builder
- Skill catalog builder
- references
- skill-catalog-builder
- outputs
- 25-skill-invocation-and-routing
- outputs
- skill-invocation-router
- Skill invocation router
- references
- skill-invocation-router
- outputs
- 26-skill-permissions-sandboxes-and-trust
- outputs
- skill-safety-reviewer
- Skill safety reviewer
- references
- skill-safety-reviewer
- outputs
- 27-skill-evals-packaging-and-portability
- outputs
- skill-release-gate
- Skill release gate
- references
- skill-release-gate
- outputs
- 28-mcp-tool-contracts-and-content
- 29-mcp-reliability-cancellation-and-flow-control
- 30-mcp-registry-supply-chain-and-drift
- 31-mcp-conformance-versioning-and-operations
- Phase 14: Agent Engineering
- 01-the-agent-loop
- 02-rewoo-plan-and-execute
- 03-reflexion-verbal-rl
- 04-tree-of-thoughts-lats
- 05-self-refine-and-critic
- 06-tool-use-and-function-calling
- 07-memory-virtual-context-memgpt
- 08-memory-blocks-sleep-time-compute
- 09-hybrid-memory-mem0
- 10-skill-libraries-voyager
- 11-planning-htn-and-evolutionary
- 12-anthropic-workflow-patterns
- 13-langgraph-stateful-graphs
- 14-autogen-actor-model
- 15-crewai-role-based-crews
- 16-openai-agents-sdk
- 17-claude-agent-sdk
- 18-agno-and-mastra-runtimes
- 19-benchmarks-swebench-gaia
- 20-benchmarks-webarena-osworld
- 21-computer-use-agents
- 22-voice-agents-pipecat-livekit
- 23-otel-genai-conventions
- 24-agent-observability-platforms
- 25-multi-agent-debate
- 26-failure-modes-agentic
- 27-prompt-injection-defense
- 28-orchestration-patterns
- 29-production-runtimes
- 30-eval-driven-agent-development
- 31-agent-workbench-why-models-fail
- 32-minimal-agent-workbench
- 33-instructions-as-executable-constraints
- 34-repo-memory-and-state
- 35-initialization-scripts
- 36-scope-contracts
- 37-runtime-feedback-loops
- 38-verification-gates
- 39-reviewer-agent
- 40-multi-session-handoff
- 41-workbench-for-real-repos
- 42-agent-workbench-capstone
- 43-frame-the-task-before-code
- Phase 15: Autonomous Systems
- 01-long-horizon-agents
- 02-star-family-reasoning
- 03-alphaevolve-evolutionary-coding
- 04-darwin-godel-machine
- 05-ai-scientist-v2
- 06-automated-alignment-research
- 07-recursive-self-improvement
- 08-bounded-self-improvement
- 09-coding-agent-landscape
- 10-claude-code-permission-modes
- 11-browser-agents
- 12-durable-execution
- 13-cost-governors
- 14-kill-switches-canaries
- 15-propose-then-commit
- 16-checkpoints-rollback
- 17-constitutional-ai
- 18-llama-guard
- 19-anthropic-rsp
- 20-openai-preparedness-deepmind-fsf
- 21-metr-external-evaluation
- 22-cais-caisi-societal-risk
- Phase 16: Multi-Agent & Swarms
- 01-why-multi-agent
- 02-fipa-acl-heritage
- 03-communication-protocols
- 04-primitive-model
- 05-supervisor-orchestrator-pattern
- 06-hierarchical-architecture
- 07-society-of-mind-debate
- 08-role-specialization
- 09-parallel-swarm-networks
- 10-group-chat-speaker-selection
- 11-handoffs-and-routines
- 12-a2a-protocol
- 13-shared-memory-blackboard
- 14-consensus-and-bft
- 15-voting-debate-topology
- 16-negotiation-bargaining
- 17-generative-agents-simulation
- 18-theory-of-mind-coordination
- 19-swarm-optimization-pso-aco
- 20-marl-maddpg-qmix-mappo
- 21-agent-economies
- 22-production-scaling-queues-checkpoints
- 23-failure-modes-mast-groupthink
- 24-evaluation-coordination-benchmarks
- 25-case-studies-2026-sota
- Phase 17: Infrastructure & Production
- 01-managed-llm-platforms
- 02-inference-platform-economics
- 03-gpu-autoscaling-kubernetes
- 04-vllm-serving-internals
- 05-eagle3-speculative-decoding
- 06-sglang-radixattention
- 07-tensorrt-llm-blackwell
- 08-inference-metrics-goodput
- 09-production-quantization
- 10-cold-start-mitigation
- 11-multi-region-kv-locality
- 12-edge-inference
- 13-llm-observability
- 14-prompt-semantic-caching
- 15-batch-apis
- 16-model-routing
- 17-disaggregated-prefill-decode
- 18-vllm-production-stack-lmcache
- 19-ai-gateways
- 20-shadow-canary-progressive
- 21-ab-testing-llm-features
- 22-load-testing-llm-apis
- 23-sre-for-ai
- 24-chaos-engineering-llm
- 25-security-secrets-audit
- 26-compliance-frameworks
- 27-finops-llms
- 28-self-hosted-serving-selection
- Phase 18: Ethics, Safety & Alignment
- 01-instruction-following-alignment-signal
- 02-reward-hacking-goodhart
- 03-direct-preference-optimization-family
- 04-sycophancy-rlhf-amplification
- 05-constitutional-ai-rlaif
- 06-mesa-optimization-deceptive-alignment
- 07-sleeper-agents-persistent-deception
- 08-in-context-scheming-frontier-models
- 09-alignment-faking
- 10-ai-control-subversion
- 11-scalable-oversight-weak-to-strong
- 12-red-teaming-pair-automated-attacks
- 13-many-shot-jailbreaking
- 14-ascii-art-visual-jailbreaks
- 15-indirect-prompt-injection
- 16-red-team-tooling-garak-llamaguard-pyrit
- 17-wmdp-dual-use-evaluation
- 18-frontier-safety-frameworks-rsp-pf-fsf
- 19-model-welfare-research
- 20-bias-representational-harm
- 21-fairness-criteria-group-individual-counterfactual
- 22-differential-privacy-for-llms
- 23-watermarking-synthid-stable-signature-c2pa
- 24-regulatory-frameworks-eu-us-uk-korea
- 25-echoleak-cves-for-ai
- 26-model-system-dataset-cards
- 27-data-provenance-training-governance
- 28-alignment-research-ecosystem
- 29-moderation-systems-openai-perspective-llamaguard
- 30-dual-use-risk-cyber-bio-chem-nuclear
- Phase 19: Capstone Projects
- 01-terminal-native-coding-agent
- 02-rag-over-codebase
- 03-realtime-voice-assistant
- 04-multimodal-document-qa
- 05-autonomous-research-agent
- 06-devops-troubleshooting-agent
- 07-end-to-end-fine-tuning-pipeline
- 08-production-rag-chatbot
- 09-code-migration-agent
- 10-multi-agent-software-team
- 11-llm-observability-dashboard
- 12-video-understanding-pipeline
- 13-mcp-server-with-registry
- 14-speculative-decoding-server
- 15-constitutional-safety-harness
- 16-github-issue-to-pr-agent
- 17-personal-ai-tutor
- 46-gradient-accumulation
- 47-checkpoint-save-resume
- 48-distributed-fsdp-ddp
- 49-lm-eval-harness
- 82-jailbreak-taxonomy
- outputs
- 83-prompt-injection-detector
- outputs
- 84-refusal-evaluation
- outputs
- 85-content-classifier-integration
- 86-constitutional-rules-engine
- outputs
- 87-end-to-end-safety-gate
- outputs
- Phase 0: Setup & Tooling
开始学习 → Check Understanding