LLM Inference Framework

What if a 125B-parameter AI model could run locally on a gaming GPU with just 12 GB of VRAM? Strata makes it possible. This book takes you inside the techniques that make it work, from MoE sparsity and memory tiers to PCIe pipelining and speculative decoding, then shows you how to install, tune and troubleshoot it yourself.

Inference engineering is now its own discipline, with unique challenges around speed, cost, reliability and scale. This practical guide takes you from core concepts to production systems, showing how modern inference architectures work, how to optimize them and how to make better engineering decisions through measurement and trade-off analysis.

What happens when AI moves beyond answering questions and begins to pursue goals?Agentic AI: The Rise of Autonomous, Goal-Driven Intelligence explores the emerging world of intelligent systems that can perceive, reason, plan, learn, collaborate, and act within defined environments.From the foundations of autonomous agents and multi-agent systems to robotics, finance, logistics, security, ethics, and future research, this book provides a structured journey through the rapidly evolving landscape of Agentic AI.Discover how goal-driven intelligence works, how autonomous agents make decisions, how multiple agents collaborate, and why safety, trust, accountability, and human oversight are essential to the future of autonomous systems.Understand the agents. Explore the architectures. Examine the applications. Think about the future.

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.

This book shows you how.What you'll be able to do:Say what one resolved ticket, generated report, or active customer actually costs youFind where the money goes in a real system, in about an hourCut your bill by 40% to 70% without making the product worse, and prove it didn'tStop a code change from tripling your costs before it shipsWalk into a budget meeting with a number and an argument instead of an apologyWhat's inside: sixteen chapters. Measure what you're paying for. Apply six levers, ordered by how much they pay back. Then keep the savings from creeping away, which is the part that decides whether any of this survives the year.

Every sentence an AI writes can carry a hidden signature, and the world's biggest models are already signing their output. AI Watermarking shows you exactly how it works, using a four-word vocabulary and arithmetic you can check with a pocket calculator. By the end you will know what a watermark can prove, what destroys it in seconds, and why short text can never be tested.

Take a practical journey through SGLang, from its core architecture to real-world deployment and optimization. Built for engineers who need to run LLM inference reliably at scale, this handbook breaks down the tools, techniques and operational know-how needed to build and maintain production-ready SGLang systems.