دانلود کتاب تسلط بر طراحی سیستم هوش مصنوعی

This book offers a comprehensive guide to designing, building, and deploying advanced AI systems for real-world applications. It bridges foundational AI concepts with practical industry implementation, covering everything from system design principles and prompt engineering for generative AI to specialized architectures such as Retrieval-Augmented Generation (RAG) and multi-modal transformers. Readers gain hands-on guidance in fine-tuning Large Language Models (LLMs), building domain-specific models, and implementing both supervised and unsupervised learning systems across diverse applications such as finance, e-commerce, healthcare, and edge computing. Through a combination of theoretical insights, practical workflows, and case studies, the book equips readers to handle challenges in data processing, model evaluation, explainability, ethical AI, and deployment strategies. From recommendation engines and sequenceto- sequence systems to custom LLMs and multi-modal AI, it provides a holistic blueprint for creating scalable, efficient, and impactful AI solutions across industries. Chapter 1: Introduction to AI System Design This chapter establishes the essential principles, and addresses the key challenges of AI system design, serving as a bridge between academic learning and real-world industry expectations. It introduces a structured framework to help readers effectively approach and solve AI system design problems. Chapter 2: Crafting Intelligent Systems Using Prompt Engineering This chapter focuses on the transformative potential of Generative AI, guiding readers through prompting techniques and frameworks such as Hugging Face and OpenAI APIs. Through real-world case studies, it demonstrates how to design intelligent systems using large language models, and equipping readers with practical skills for building advanced AI solutions. Chapter 3: Developing Retrieval-Augmented Generation Systems This chapter demystifies RAG architectures that integrate information retrieval with generative models. It provides detailed guidance on database selection, explores toolkits such as LangChain and Llama Index, and discusses deployment trade-offs. The concepts are reinforced through case studies and interview scenarios, offering readers practical insights into real-world applications. Chapter 4: Enhancing Systems Through LLM Finetuning This chapter provides readers with hands-on experience in advanced fine-tuning methods such as PEFT, LoRA, and instruction tuning. It bridges these fine-tuning strategies to real-world business applications, demonstrating how to adapt large models for specialized tasks and practical deployment scenarios. Chapter 5: Designing Financial Risk Prediction Systems Using Supervised Learning This chapter presents a comprehensive, end-to-end journey in designing supervised machine learning systems for financial risk analysis. It guides readers through vital stages such as data processing, model evaluation, explainability, and privacy considerations. The chapter concludes with practical deployment strategies and domainspecific case studies, offering actionable insights for real-world applications. Chapter 6: Implementing Unsupervised Learning Systems In this chapter, readers explore unsupervised learning methods, including clustering and topic modeling. The discussion emphasizes embedding strategies, evaluation techniques, and practical CRM applications for customer segmentation and actionable insights. Chapter 7: Building Recommendation Systems for E-Commerce This chapter traces the evolution of recommendation systems, guiding readers from collaborative filtering approaches to advanced graphbased models. It places special emphasis on ethical AI, explainability, experimentation, and effective deployment strategies, providing a comprehensive perspective on building modern recommendation engines. Chapter 8: Building Image Classification Models for Edge Devices This chapter teaches readers how to design optimized computer vision systems specifically for resource-constrained environments. It covers essential topics such as data annotation, model architecture, optimization techniques, and deployment strategies for mobile and edge devices, all illustrated with real-life case examples. Chapter 9: Designing Sequence-to-Sequence Systems This chapter examines sequence-to-sequence models that drive applications such as translation, summarization, and speech recognition. It guides readers through key topics including data challenges, LSTMs, transformer-based design, and performance evaluation, all illustrated with a practical summarization case study. Chapter 10: Building Domain-Specific LLMs from Scratch This chapter offers an in-depth exploration of building customized LLMs, guiding readers through data curation, transformer architecture, distributed training, and evaluation. It presents a comprehensive blueprint for developing specialized LLMs, exclusively tailored to meet the industry requirements. Chapter 11: Building Multimodal Applications for Healthcare The final chapter delves into multi-modal AI architecture, guiding readers through dataset preparation, architectural alignment, and deployment trade-offs. It culminates with a comprehensive healthcare case study and an interview guide.