Machine Learning

What really happens inside a neural network? This book explores the tools and ideas researchers use to uncover how AI systems work, from foundational concepts to practical experiments. Learn to separate solid findings from bold claims, develop useful research skills and prepare for a field that is changing as quickly as the technology itself.

From your first predictive model to the ensembles that win competitions — the machine-learning core of the Data Science, End to End series, built from first principles in clean, idiomatic Python.

Your LLM pipeline passed every check, and a fallback model still served 50 of its 71 successful calls. This case study follows one real multi-stage pipeline through its failures and shows how to make each stage traceable, attributable and honestly measured.

Decision models answer a typed question with a calibrated probability in one forward pass: no token stream, no parsing, no values outside your schema. This guide explains TypeSafe's Jev and the open "JEV-like" family, then shows how to choose, run, serve, fine-tune and evaluate local variants on your own hardware.

Your Sharpe ratio is 2.8. Your model is wrong.Not because the code is broken, but because somewhere between the raw data and the equity curve, the future leaked into the past. A centered rolling window. A normalization fit on the full sample. A fundamentals join on the wrong date. Each one runs clean, and each one produces a number that disappears the week real money replaces history.Machine Learning for Market Prediction is a quant reviewer’s playbook for building a pipeline you can actually trust: point in time data, leakage-free features, triple barrier labels, embargoed walk-forward validation, realistic costs, and SHAP audits that trace a suspicious feature to its root cause.It will not make your model more profitable. It will tell you the truth about it.Distrust your backtest. Then prove it right.

Build the engineering foundations required to design and develop production-grade AI systems—from Python and FastAPI to databases, security, semantic search, RAG, LLM architecture, and AI application performance.Go beyond Python programming and start thinking like an AI Engineer.The AI Engineer — Volume I takes you from professional Python engineering to the foundations of modern AI application development.You will build the engineering knowledge required to create reliable AI systems—not just experimental notebooks or simple API calls.A project-based AI engineering book covering Python production engineering, FastAPI, PostgreSQL, authentication, testing, reliability, embeddings, semantic search, RAG, LLM architecture, AI pipelines, caching, and performance optimization.AI EngineeringAI EngineerArtificial IntelligenceGenerative AILLMLarge Language ModelsPythonFastAPIRAGRetrieval Augmented GenerationSemantic SearchEmbeddingsVector SearchLLM ApplicationsAI SystemsMachine Learning EngineeringBackend EngineeringProduction AIAI ArchitectureSoftware Engineering

사내 LLM 게이트웨이를 설계하고 운영하는 실무서입니다. 키리스 인증과 6단계 RBAC, 토큰 쿼터·예산과 비용 통제, 엔터프라이즈 규모 MCP, 운영 런북까지 다룹니다. 영문판 The AI Gateway Playbook의 한국어판(PDF 75쪽)입니다.

A practitioner's playbook for building an in-house LLM gateway: keyless auth, six-tier RBAC, quotas and budgets, per-token cost control, MCP at enterprise scale, and the operations to run it all. About 32,000 words with diagrams, configs, and checklists.

The ultimate bridge from high school AP Physics to modern AI. Start with the familiar physics you know, translate it into Python, and use it as a launchpad to build advanced Machine Learning models and neural networks. Includes 100% solved exercises and click-to-run Google Colab labs.

Build one real AI system end to end, from raw job postings to a deployed, tested API, and learn production AI engineering on the way. 24 chapters, runnable code, and numbers you can reproduce.

机器人真正学会做饭，不是因为工程师写完了每一步，而是因为它能从身体经验、失败与恢复中形成自己的手艺。

为什么参数不大的模型，把信息放对地方就能反超更大的模型？本书以"压缩"为主线，把大模型的表示、泛化、推理、RAG 与智能体统一成一道信息管理题：什么该压、什么不能丢、什么该外置、放在哪里、何时取回。18 章、190 张原创图解，从熵一路讲到世界模型。

Go beyond calling transformer libraries and learn what actually happens under the hood. Starting from first principles, this book builds transformers from scratch with NumPy and PyTorch, then takes you through training, optimization and production inference. Along the way, you’ll work through the math, code and practical techniques behind modern transformer systems.

The intelligence is in the data. Learn to build AI systems that answer with grounded confidence, not fluent invention — through vector databases, RAG, and semantic search, from first principles to production-ready blueprints.

We'll be working on a platform made up of seventy-nine recipes together. It starts off as a simple task, printing a line, but by the last chapter it covers extraction, warehousing, containers, machine learning and incident response. You can't just throw away examples in your work, and you shouldn't be doing that in your examples either. You don't need to be an Airflow expert to get started. What you're really learning here isn't a tool. It's all about making sure work is repeatable, observable and safe to rerun.