Recommended Reading

Curated resources for going deeper on agentic engineering topics. This list will be updated as the field evolves.

Books

Foundations

Foundational books on AI, machine learning systems, and data architecture.

• Artificial Intelligence: A Modern Approach (4th ed.) – Stuart Russell & Peter Norvig (2020). The authoritative AI textbook covering intelligent agent principles, search algorithms, reasoning, and machine learning fundamentals.
• Software Engineering at Google – Titus Winters, Tom Manshreck & Hyrum Wright (2020). Practical engineering practices for sustaining large codebases—highly relevant for teams adopting AI-assisted workflows.
• The Pragmatic Programmer (20th Anniversary Edition) – Andrew Hunt & David Thomas (2019). Timeless advice on writing clean, flexible code that remains vital even as AI tools enter the mix.
• Designing Machine Learning Systems – Chip Huyen (2022). Holistic guide to building reliable ML-powered applications, from data processing to deployment and monitoring.
• AI Engineering – Chip Huyen (2025). Practical framework for building applications with foundation models, bridging traditional engineering and modern AI development.
• Designing Data-Intensive Applications – Martin Kleppmann (2017). Deep dive into reliable, scalable systems—crucial architectural knowledge before layering on AI features.

Manifestos and philosophy

Frameworks for thinking about AI-assisted development at a higher level.

• The o16g Manifesto – Cory Ondrejka (2025). “Outcome Engineering” — 16 principles for reorienting development around outcomes rather than code. Argues for managing to cost (tokens) instead of capacity (engineer-hours), measuring success by verified impact, and treating the backlog as a relic of human limitation. From the CTO of Onebrief, co-creator of Second Life, and former engineering leader at Google and Meta.

Context engineering

The emerging discipline of providing agents with the right information at the right time.

• Context Engineering for AI Agents – Tobi Lutke - Shopify CEO on why context engineering is the new skill
• Context Engineering – Andrej Karpathy - “Prompt engineering is dead, context engineering is king”
• AGENTS.md - Open format for persistent agent context, used by 60k+ open-source projects

Prompt engineering

Guides to effective prompting and AI interaction.

• Google Cloud Prompt Engineering Guide – Comprehensive official guide covering prompt format, examples, multi-turn interactions, and best practices.
• DAIR Prompt Engineering Guide – Mostafa Samir et al. Extensive open-source guide aggregating the latest techniques, from basic design to advanced strategies like multi-step reasoning and tool use.
• Learn Prompting – Sander Schulhoff et al. (2024). Free course covering fundamentals to advanced techniques, used by 3M+ users including Fortune 500 teams.
• The Ultimate Guide to Prompt Engineering – Lakera (2025). Modern best practices with focus on real-world usage and security, including defense against prompt injections.
• Prompt Engineering for Generative AI – James Phoenix & Mike Taylor (2023). Practical book on principles and patterns across domains—NLP, image generation, and code generation.
• Prompt Engineering for LLMs – John Berryman & Albert Ziegler (2024). Advanced strategies from GitHub Copilot developers covering token management, few-shot prompting, and workflow patterns.

Software engineering

Classic and modern software engineering texts relevant to AI-assisted development.

• Clean Code – Robert C. Martin (2008). Classic manual on maintainable code—crucial for recognizing and refactoring AI-generated code into well-structured designs.
• Refactoring (2nd ed.) – Martin Fowler (2018). Seminal guide to systematically restructuring code, invaluable for continuously improving AI-written code.
• Accelerate – Nicole Forsgren, Jez Humble & Gene Kim (2018). Data-driven research on high-performing teams, introducing DORA metrics—essential baseline when integrating AI into workflows.
• Agentic AI Engineering – Yi Zhou (2025). Forward-looking guide reframing software engineering for AI agents, covering the Agentic Stack and maturity models for scaling agents to production.
• The LLM Engineering Handbook – Paul Iusztin & Maxime Labonne (2024). Operations guide covering prompt engineering, fine-tuning, RAG, and patterns for putting LLMs into production.
• A Philosophy of Software Design (2nd ed.) – John Ousterhout (2021). Concise essays on managing complexity—lessons that complement AI tools by helping developers shape architecture and keep complexity under control.

Research papers

Key papers on agents

Academic papers defining agentic systems and patterns.

• ReAct: Synergizing Reasoning and Acting in Language Models – Yao et al. (2022). Introduced the paradigm of interleaving reasoning traces with actions, enabling LLMs to gather information and adjust plans mid-stream. Foundation for many agent frameworks.
• HuggingGPT: Solving AI Tasks with ChatGPT and its Friends – Shen et al. (2023). LLM-powered controller that orchestrates specialized models for complex multi-modal tasks, using natural language as the glue between tools.
• Toolformer: Language Models Can Teach Themselves to Use Tools – Schick et al. (2023). Meta AI’s method for LLMs to self-train on API usage, learning when and how to call external tools strategically.
• Generative Agents: Interactive Simulacra of Human Behavior – Park et al. (2023). Architecture for believable simulated agents with long-term memory and planning, demonstrating human-like behavior over extended periods.
• Voyager: An Open-Ended Embodied Agent with Large Language Models – Wang et al. (2023). First LLM-driven lifelong learning agent in Minecraft, continuously exploring and accumulating skills without human intervention.
• Reflexion: Language Agents with Verbal Reinforcement Learning – Shinn et al. (2023). Framework enabling agents to learn from mistakes through self-reflection in natural language, without formal fine-tuning.

Language models for code

Research on code generation, understanding, and synthesis.

• StarCoder: may the source be with you! - Open-source Code LLM with 8K context and infilling capabilities

Human-AI collaboration

Studies on how humans and AI systems work together effectively.

• Experimental Evidence on the Productivity Effects of Generative AI – Noy & Zhang (2023, Science). Landmark study showing ChatGPT users completed tasks ~40% faster with 18% higher quality—lower-skilled participants benefited most.
• The Productivity Effects of Generative AI: Evidence from GitHub Copilot – Cui et al. (2024). Field experiment at Microsoft/Accenture showing 12–22% more PRs completed per week with Copilot access.
• When Humans and AI Work Best Together – MIT Sloan (2025). Meta-analysis finding collaboration shines when humans are individually better than AI—success requires calibrating when to trust AI.
• Coding on Copilot: Data Suggests Downward Pressure on Code Quality – GitClear (2023). Analysis of 153M lines showing code churn doubled with AI use—teams need practices to keep quality in check.
• GitHub Copilot: Asset or Liability? – Moradi Dakhel et al. (2023). Copilot valuable for experts who can vet output, but potential liability for novices who accept faulty suggestions.

Online resources

Blogs and newsletters

Regular sources of insight on AI and development.

• 2025: The year in LLMs – Simon Willison - Comprehensive annual review of LLM developments

Podcasts

Audio content covering agentic development.

• The RPI workflow - Build Wiz AI Show (Podcast) - Audio discussion on advanced AI coding

Video content

Tutorials, talks, and demonstrations.

• The Complete AI Coding Course (2025) - Hands-on course covering Cursor and Claude Code

Communities

Places to connect with others working in this space.

• Kilo Code Discord - Our community for discussing agentic engineering
• GitHub Discussions - Longer-form conversations about this guide

---

Contributing resources

Know a great resource we should include? This list grows through community contributions.

How to add a resource:

1. Open a PR with your addition
2. Include a brief description of why it’s valuable
3. Place it in the appropriate section

We’re especially looking for:

• Books and papers that shaped your understanding
• Tutorials that actually helped you get started
• Communities where you’ve found good discussions
• Case studies from real implementations

Join the conversation on Discord if you want to discuss what should be included.

---

This list is actively maintained by the community. Your recommendations help others learn faster.