Favourite Books

Jay Alammar's visual essays have taught more engineers to think clearly about transformers than any university course. This book extends that tradition – how tokenisation works, what attention actually does, how embeddings carry meaning. Practical, visual, and directly applicable to anyone building with LLMs.

You only truly understand something when you can build it. Raschka walks through implementing a GPT-like model from first principles – attention mechanisms, context windows, training loops. The architecture stops being a black box. The moment context windows stop feeling mysterious is worth the read alone.

Prompting well is a genuine skill with real structure. This is the most practical book on it – covering patterns, chain-of-thought techniques, and prompt design across different models and use cases. What you ask matters. How you ask it matters more.

How to inject domain knowledge into AI models without retraining them – the full retrieval pipeline from embeddings to vector databases to chunking strategies. Essential for anyone building AI systems that need to know things beyond their training cutoff.

The most comprehensive treatment of building AI systems end-to-end. Covers infrastructure, APIs, tool use, evaluation frameworks, gateways, agents, and production deployment. If you read one book about how the modern AI stack fits together as a whole – this is it.

The craft of keeping AI models working reliably after deployment – data pipelines, feature stores, model monitoring, serving infrastructure, and avoiding silent failures in production. The gap between a prototype and a production system lives in this book.

The most thoughtful examination of what it means to work alongside AI – not just use it. Mollick, a Wharton professor who studies AI adoption closely, shows how to treat AI as a genuine collaborator. Essential for understanding what the agentic era actually demands from humans.

The clearest account of why teaching AI systems to reliably do what we want is harder than it appears – and why the gap between a model’s behaviour and human intent can be catastrophic at scale. Essential reading before you deploy AI systems that make real decisions.

One of the world’s leading AI researchers makes the case for building systems genuinely aligned with human values – and explains why current approaches have a structural flaw. The intellectual foundation for anyone serious about responsible AI thinking.

DeepMind’s co-founder charts what happens as AI and synthetic biology converge – not a doom scenario, but an honest mapping of the decisions governments, companies, and engineers need to make now. The strategic overview that puts every architecture decision in context.

Traces the real cost of AI systems – the energy, the data labour, the supply chains, the power structures. For architects deploying AI at enterprise scale, understanding these dimensions is part of responsible system design, not a separate conversation.

How non-conformists move the world. Reinforced my belief that the most valuable ideas are the ones that feel uncomfortable at first.

A humbling lens on human history. Puts every career milestone, every architectural decision, every ambition into breathtaking perspective.

The science of tiny changes and compound improvement. The foundation of how I build discipline – in engineering, in learning, in life.

The bible of professional software craftsmanship. I return to its principles every time I review code or mentor an engineer.

A powerful Pakistani narrative connecting personal resilience with national identity. It speaks to the grit in my roots.

Reframed how I think about leadership and legacy. We are not playing to win – we are playing to keep playing and keep improving.

The Japanese art of finding purpose. This book helped me consciously define what happiness means for my family and myself.

A vivid case for owning your mornings. My daily rituals, continuous learning habit, and early-hour focus all owe something to this book.

Timeless wisdom on the art of professional programming. Shaped my engineering instincts more than any framework or tool ever has.

A practical guide to the skills that matter most in leadership. High EQ, I believe, matters more than high IQ in almost every real-world situation.

A sobering reminder that success is rarely as deliberate as we think. Humility about luck is the first step toward genuine wisdom – in markets, in careers, in life.

The book that revealed the invisible architecture of decision-making. System 1 vs System 2 thinking changed how I approach technical choices, team dynamics, and risk.

A masterclass in intellectual humility. The best engineers I know share one trait: they are willing to be wrong. This book explains why that's so rare – and how to cultivate it.

Why good companies fail when disruption arrives. Essential reading for any cloud architect navigating legacy systems and the relentless pace of technology change.

Decision quality is not the same as outcome quality. This reframe changed how I evaluate architectural decisions – especially the ones that don't go as planned.

How to be genuinely less wrong over time. The calibration techniques here directly informed how I think about risk, estimation, and uncertainty in cloud architecture.

Smart people make systematic errors – and intelligence can actually amplify those errors. A necessary corrective for anyone who considers themselves an expert.

Talent is overrated. Sustained passion and perseverance – applied consistently over time – produce more extraordinary outcomes than any natural gift. A lesson Pakistan taught me early.

"Follow your passion" is terrible advice. Rare and valuable skills create rare and valuable careers. The craftsman mindset, articulated perfectly.

The companion to Clean Code that operates at the system level. Separation of concerns, dependency rules, and architecture boundaries – principles I apply every day in cloud system design.

Before strategy, before roadmaps, before architecture – ask why. This book gave me language for something I always felt: purpose-led work produces better systems and better people.

A childhood fascination that never left. Engineering precision, aerodynamic mastery, and the relentless pursuit of performance – lessons that apply as much to cloud systems as to aircraft design.

A masterclass in data literacy and fighting our instinct to catastrophize. Rosling's ten instincts that distort our worldview belong on every engineer's and leader's reading list.

Counterintuitive and sharp. Choose what genuinely matters and protect your energy accordingly. One of the most honest self-help books ever written.

The career map I wish I had when I started. From junior to staff engineer – practical, honest, and grounded in how real engineering careers actually progress.

More than interview prep – it sharpened my problem-solving instincts and reminded me that clarity of thought under pressure is a craftsman's core skill.

A fable about what truly matters. Read it twice – once for the story, once for the philosophy. A reminder that the greatest architecture you can build is the architecture of your own life.