Claude Code Is Not Your Architect
Claude Code can accelerate software delivery, but it cannot replace architectural judgment. Once an AI coding agent can understand a repository, edit files, run commands, create pull requests, and interact with tools, the real question is no longer “Can it generate code?” The real question is, “Can we safely govern what is allowed to change, test, approve, deploy, and explain?”
This talk introduces a practical architecture and governance model for using Claude Code in real software teams. Participants will learn how to move beyond prompt-to-code experiments and build safer workflows that incorporate context, limits, architectural rules, human approval, delivery gates, and evidence trails.
AI coding agents are changing software delivery. Claude Code can help developers understand codebases, modify files, run commands, generate tests, support Git workflows, and accelerate pull request creation. That speed is powerful, but it also introduces a new architecture risk: generated code can look correct while hiding security gaps, weak tests, unsafe retries, broken authorization, poor observability, excessive cloud permissions, and fragile production behavior.
This talk introduces the role of the Claude Code Architect: not as a replacement for developers, but as the person who defines how agentic coding is safely used across repositories, teams, CI/CD pipelines, and cloud systems. We will cover project memory, CLAUDE.md, plan-first workflows, permission boundaries, sandboxing, MCP/tool governance, pull-request policies, secure code review, CI/CD gates, observability, and production-readiness checks.
The goal is simple: use Claude Code to move faster without surrendering architectural control.
Main audience promise
By the end of this talk, participants will understand how to safely introduce Claude Code into a real engineering workflow without creating fragile, insecure, unreviewable, or ungoverned production systems.
⸻
Why this talk matters now
AI coding is moving from simple autocomplete to agentic software delivery. The new generation of tools can inspect repositories, plan changes, edit files, run commands, participate in pull requests, and connect to external tools through protocols such as MCP. That means AI coding agents are becoming active participants in the SDLC.
This creates new questions for architects and senior developers:
- What context should the agent be allowed to see?
- What commands should it be allowed to run?
- What files should it never modify?
- What secrets and credentials must stay outside its reach?
- What changes require human approval?
- What CI/CD gates must block unsafe output?
- What evidence should be preserved for review and audit?
- When should generated code be rejected even if it compiles?
This talk answers those questions through a practical architecture framework.
⸻
Talk details
Primary focus
This talk is not a Claude Code tutorial. It is an architecture talk about safe autonomy in software delivery.
It focuses on:
- Claude Code as an engineering workflow actor
- Risks introduced by agentic coding
- Repository-level architecture standards
- Context and project memory
- CLAUDE.md as an architecture control surface
- Plan-first workflows
- Prompt-to-plan-to-diff discipline
- Human-in-the-loop approval
- MCP and external tool governance
- CI/CD safety gates
- Secure code review
- Cloud-readiness checks
- Evidence and auditability
- Production-readiness decisions
About Rohit Bhardwaj
Rohit Bhardwaj is a Director of AI & Data Architecture at Salesforce, where he focuses on enterprise AI, agentic systems, cloud-native architecture, distributed systems, data platforms, security, and large-scale transformation.
Over his career, Rohit has designed and led complex enterprise platforms across AWS, Google Cloud, microservices, real-time data, API ecosystems, resilient distributed systems, and AI-enabled architectures. His work increasingly focuses on the challenges enterprises face as software evolves from deterministic services to AI-native and agentic systems—particularly around reliability, governance, evidence, security, observability, cost, and safe autonomy.
Rohit is the author of System Design with AI Interview Guide: Designing Scalable, Agentic, and Defensible Systems, published by Apress. The book presents a modern approach to system design covering scalability, distributed systems, AI architecture primitives, security, reliability, economics, agentic systems, and real-world architectures including e-commerce, ride sharing, payments, fraud detection, messaging, video streaming, file storage, and search. (Springer Link)
Book:
Amazon: https://a.co/d/09Zs1twa
Publisher / Springer Nature: https://link.springer.com/book/10.1007/979-8-8688-2782-2
O'Reilly: https://learning.oreilly.com/library/view/system-design-with/9798868827822/
Rohit is also an O’Reilly instructor and a frequent speaker at technology conferences including No Fluff Just Stuff, UberConf, GIDS, and other international events. His talks focus on practical architecture lessons from building and operating complex systems, including AI control planes, trusted agents, inference at scale, evidence-first RAG, AI security, distributed-system failure, and AI-era software architecture.
As a trusted advisor and architecture leader, Rohit works at the intersection of business strategy and deep technical architecture—helping teams translate complex business problems into scalable, resilient, secure, and economically sustainable systems.
Rohit holds an MBA in Corporate Entrepreneurship from Babson College and graduate-level education in Computer Science from Boston University and Harvard University.
Connect with Rohit:
LinkedIn: http://linkedin.com/in/rohit-bhardwaj-cloud
X / Twitter: @rbhardwaj1