The Platform Engineer’s Handbook • Ajay Chankramath & Kaspar von Grünberg • GOTO 2026
This conversation with Ajay Chankramath, author of 'The Platform Engineer’s Handbook,' delves into why practical, code-first guidance is essential for building Internal Developer Platforms. He argues that developer adoption failures stem from a "product discipline gap," not a technology one, emphasizing developer experience as a first-class outcome. The discussion covers the book's arc from foundations to enterprise-grade features and its focus on 100% open-source, vendor-agnostic tooling. Crucially, it highlights how agentic AI raises the stakes for platform engineering, requiring new IDP layers for agent context, memory, and guardrails, asserting that these must be built, owned, and operated internally for safe and productive AI adoption.
Ajay Chankramath’s “The Platform Engineer’s Handbook” addresses a critical gap in platform engineering literature: the lack of a hands-on, practitioner’s guide to building a real, production-grade Internal Developer Platform (IDP) from scratch. Unlike many theoretical resources, this book provides runnable steps, validated exercises, and full code, making it invaluable for those starting their platform journey.
The Product Discipline Gap and Developer Experience
Chankramath argues that the primary reason platforms fail to be adopted isn’t a technology gap, but a “product discipline gap.” This refers to a fundamental lack in the feedback loop, an insufficient understanding of users, and a failure to treat developer experience (DX) as a first-class outcome. Many organizations are enamored with new technologies but neglect the ultimate business outcome and user needs. Sustainable platform building requires focusing on intrinsic value and delivering a platform that developers truly deserve, rather than one merely easy to justify to leadership. Platform engineering is presented as a continuous investment and a blend of social and technical innovation.
Book’s Structure and Approach
The book follows a logical progression, beginning with the core foundations of a platform. It covers essential components like Kubernetes-based runtimes, service meshes, and identity management (OIDC). Readers build a functioning platform capable of running microservices. Subsequent chapters incrementally add capabilities, focusing on self-service for platform consumers through developer portals and onboarding processes. The latter part of the book addresses enterprise-grade concerns and AI readiness, including compliance-as-code (using tools like OPA and event emission controllers), FinOps, chaos engineering, and disaster recovery. A unique aspect is a single case study that evolves throughout all 14 chapters, allowing readers to see the platform’s value expand in a real-life scenario.
Industry Agnostic and Vendor Sovereignty
The book aims to be industry-agnostic, primarily benefiting enterprises, and is designed for practicing software engineers or operations professionals. A significant emphasis is placed on vendor agnosticism and sovereignty. Chankramath highlights the importance of building solutions that are not tied to specific providers, especially given geopolitical and licensing uncertainties. To achieve this, the book exclusively uses 100% open-source tools, with dedicated sections detailing their installation and configuration. This approach lays the foundation for long-term sustainability and resilience against external policy shifts.
The Transformative Impact of AI Agents on SDLC
Perhaps the most crucial aspect for a technical audience today is the book’s deep dive into the implications of AI agents. The traditional Software Development Life Cycle (SDLC) assumes human involvement at every stage. However, as AI agents evolve from mere tools to “genuine actors” capable of understanding context, implementing changes, validating output, and responding to signals, the SDLC fundamentally transforms. Chankramath identifies the need for a new layer within IDPs to accommodate agents, encompassing:
- Context and Memory: Providing agents with a comprehensive knowledge base.
- Capabilities and Tool Registry: Defining how agents execute tasks and interact with systems.
- Guardrails: Implementing mechanisms for data loss prevention, security, and observability.
He argues that getting agentic development right is primarily a platform engineering problem, not solely an AI team’s responsibility. The advent of AI significantly enhances the value of platform engineering. Crucially, the quality that makes agents safe and productive “cannot be bought off the shelf.” Instead, it “must be built, owned, and operated by you.” In the AI era, the true differentiator is not which frontier model an organization uses, but whether its platform foundations are solid enough to enable safe and productive agent deployment.
The Challenge of Practicality
Writing the book was a demanding endeavor due to the commitment to practicality over theory. Chankramath aimed to create a resource that readers could immediately implement, requiring extensive development of a full-stack, open-source solution from foundational elements to agentic AI integration. This involved distilling years of experience into actionable patterns and explicitly highlighting common pitfalls, ensuring the content is not just informative but genuinely usable.