# Vivek Chopra

> Notes on AI-native platforms, regulated markets, scaling engineering orgs, and operational reliability.

Vivek Chopra is a product-oriented engineering executive with 20+ years of experience building and scaling engineering organizations. He has led engineering across product development, infrastructure, security, compliance, and quality at companies including Soraban, Indigov, Very Good Security, Zendesk, Mashery (Intel/TIBCO), and Sony Computer Entertainment America. He has built systems serving 180+ million Americans, operated FedRAMP and SOC 2 compliant platforms, and managed infrastructure handling billions of API transactions. He holds 6 patents and has co-authored 10 books on web services and distributed systems.

## Focus areas

- AI-native platforms: Building with models in the loop: specs, guardrails, agents and the new SDLC.
- Regulated markets: Security, authorization and compliance where the work gets audited.
- Scaling engineering orgs: Culture, process, metrics and the people side of engineering leadership.
- Operational reliability: Failure rates, testing and the habits that keep production boring.

## Posts

- [Altdown: RSS for agents*](https://www.vivekchopra.com/posts/altdown): Altdown is a Markdown copy of each page, announced with one link element and listed in a JSON Feed. RSS says what is new. Altdown says what a page says. This site publishes copies at the page URL plus .md, and an index at /altdown.json.
- [Plan before you prompt](https://www.vivekchopra.com/posts/plan-before-you-prompt): How to steer coding agents with four small documents (architecture, ADR, plan, tasks) so your team makes the design decisions instead of the agent.
- [Guardrails for AI-generated code](https://www.vivekchopra.com/posts/ai-generated-code-guardrails): A four-layer guardrails model for AI-generated code, and notes on a shrinking architectural half-life.
- [Death to Scrum?](https://www.vivekchopra.com/posts/death-to-scrum): If the economics of software development are changing this dramatically, what happens to Scrum? Covers what the Agile Manifesto got right, what Scrum was actually for, how agents change the underlying conditions, a proposed new set of values, practical steps for gradual evolution, and which metrics survive the transition.
- [The New SDLC](https://www.vivekchopra.com/posts/new-sdlc): How AI is compressing the development cycle and changing how product, engineering, design, QA, and security teams operate. Covers spec-driven development, AI-assisted code review, automated test generation, and the emerging threat surface introduced by AI agents.
- [From Reading Changelogs to Finding Zero-Days: LLMs and the Future of AppSec](https://www.vivekchopra.com/posts/llm-appsec): The trajectory from LLMs reading shadow-patched changelogs at BSides SF 2025 to Claude Mythos Preview finding thousands of zero-days in every major OS and browser. Covers the dual-use problem and why the old model of security patches cannot survive contact with systems that operate at machine speed.
- [Spec First, Then Code](https://www.vivekchopra.com/posts/spec-driven-development): LLMs have no context — you have to give it to them. Covers AGENTS.md and CLAUDE.md for persistent project context, spec-driven development with GitHub Spec Kit and BMAD, and screenshots as a tool for precise visual communication.
- [Something Has Changed With AI Coding](https://www.vivekchopra.com/posts/ai-coding-changed): On context engineering, AGENTS.md, spec-driven development, and what Claude Opus 4.6 represents as a qualitative shift in AI coding capability. References the ETH Zurich research on LLM-generated vs human-written context files and Professor Knuth's paper on Claude.
- [OAuth Is Not Enough: Authorization Challenges for Autonomous AI Agents](https://www.vivekchopra.com/posts/ai-agent-auth): Why OAuth 2.0 is insufficient for autonomous AI agents, and a proposed layered mitigation strategy covering task-scoped tokens, policy engines, and audit hooks. Includes the full working paper published on TechRxiv.
- [Managing by Metrics](https://www.vivekchopra.com/posts/managing-by-metrics): How to select and use engineering metrics. Covers vanity metrics, watermelon metrics, change failure rate, leading vs. lagging indicators, and a worked example metric set including value shipped, KTLO ratio, bug impact score, support SLIs, and uptime.
- [Engineering Culture and Principles](https://www.vivekchopra.com/posts/engineering-culture): A working document written for an engineering team, defining culture through ownership, agency, and customer empathy, and the principles that guide how to design, build, and operate software. Includes a downloadable PDF.
- [Giving Negative Feedback](https://www.vivekchopra.com/posts/giving-negative-feedback): A practical framework for delivering negative feedback effectively, including the Toastmasters what-you-saw/heard/felt template and how to phrase feedback as encouragement to improve rather than criticism.
- [Parsing Text Files in Shell](https://www.vivekchopra.com/posts/parsing-text-files-shell): How to use IFS to parse delimited files in bash, with worked examples and notes on sparse array gotchas.
- [9 Years of Scrum, and All I Got Is a Lousy T-shirt (Sizing)](https://www.vivekchopra.com/posts/nine-years-of-scrum): Lessons from nine years of scrum including when it does not fit, how to inspect and adapt, standups, metrics, the broken windows effect, and the importance of big-picture architectural thinking.

## Contact and Links

- Website: https://www.vivekchopra.com
- LinkedIn: https://www.linkedin.com/in/chopra
- GitHub: https://github.com/vivekchopra
- RSS Feed: https://www.vivekchopra.com/feed.xml
- Sitemap: https://www.vivekchopra.com/sitemap.xml