Chapter 15 References

Books

• Ian Goodfellow, Yoshua Bengio, and Aaron Courville, Deep Learning, MIT Press, 2016.
• Chip Huyen, Designing Machine Learning Systems, second edition, O’Reilly Media, 2024.
• Chip Huyen, AI Engineering: Building Applications with Foundation Models, O’Reilly Media, 2025.
• Daniel J. Jurafsky and James H. Martin, Speech and Language Processing, draft third edition, 2024.

Websites

• ScaNN: Scalable Nearest Neighbors — Google Research implementation and algorithm notes for partition-and-search, quantization, and rescoring.
• Accelerating Large-Scale Inference with Anisotropic Vector Quantization — the ScaNN paper.
• DiskANN: Fast Accurate Billion-point Nearest Neighbor Search on a Single Node — the DiskANN and Vamana system paper.
• DiskANN source code — Microsoft implementations of out-of-core and in-memory vector indexing.
• Efficient and Robust Approximate Nearest Neighbor Search Using HNSW — the HNSW paper.
• FAISS wiki — implementation notes for exact and approximate vector indexes, including IVF and product quantization.
• Pinecone documentation — managed vector database concepts and API guidance.
• Qdrant documentation — self-hosted vector search, collections, and payload filtering.
• Milvus documentation — distributed vector database architecture and index guidance.
• LangChain RAG tutorials — retrieval, indexing, and retrieval-chain patterns.
• LlamaIndex documentation — data connectors, indexes, query engines, and agent workflows.
• DeepSpeed documentation — ZeRO, pipeline parallelism, and large-scale training configuration.
• Megatron-LM source code — NVIDIA implementation of tensor, pipeline, sequence, and context parallelism.
• PyTorch FSDP documentation — explicit parameter, gradient, and optimizer-state sharding.
• vLLM documentation — distributed LLM inference, scheduling, paged KV caches, and supported optimizations.
• Efficient Memory Management for Large Language Model Serving with PagedAttention — the PagedAttention and vLLM systems paper.
• Orca: A Distributed Serving System for Transformer-Based Generative Models — the continuous-batching serving design.
• LangChain agents — tool-calling agent loops and runtime state.
• LlamaIndex agents — prebuilt agents and event-driven agent workflows.
• ReAct: Synergizing Reasoning and Acting in Language Models — the interleaved reasoning-and-action paper.
• Hugging Face PEFT documentation — adapter configuration and PEFT integration.
• LoRA: Low-Rank Adaptation of Large Language Models — the LoRA paper.
• QLoRA: Efficient Finetuning of Quantized LLMs — the QLoRA paper.
• Deep Reinforcement Learning from Human Preferences — the foundational RLHF paper.
• Direct Preference Optimization: Your Language Model is Secretly a Reward Model — the DPO paper.
• Hugging Face TRL documentation — supervised, reward, DPO, and reinforcement-learning trainers.