Architecture Of Claude Code - Reverse Engineered
Hosted by Vikash Rungta
.png&w=1536&q=75)
Sun, Mar 8, 2026
4:00 PM UTC (1 hour)
Virtual (Zoom)
Free to join
Go deeper with a course
Zero to Launch-GenAI Product Builder Bootcamp
Vikash Rungta
Expert AI PM who built GenAI products at Meta and top global tech firms.
Sun, Mar 8, 2026
4:00 PM UTC (1 hour)
Virtual (Zoom)
Free to join
Go deeper with a course
Zero to Launch-GenAI Product Builder Bootcamp
Vikash Rungta
Expert AI PM who built GenAI products at Meta and top global tech firms.
What you'll learn
The Paradigm Shift
Why DAGs fail and why the "Harness" is the new body for the AI brain.
The 6 Core Patterns
From "Primitives > Integrations" to "Context as a Scarce Resource."
Failure Modes
How to architect specifically against Runaway Loops, Amnesia, and Permission Roulette.
Steal This
What you should copy immediately for your own internal or customer-facing agents.
Why this topic matters
We are shifting from workflows (n8n, LangChain) to true autonomous agents. I reverse-engineered Claude Code to show the 6 architectural patterns enabling this shift. Most “AI agents” are glorified workflows—brittle code driving a model. True autonomy requires moving from code controlling the model to the model controlling the loop. This session examines Claude Code’s “System 2” architecture.
You'll learn from
Vikash Rungta
xPM Meta SuperIntelligence Labs, Instructor GenAI @Stanford Continuing Studies
- Product Manager, Meta SuperIntelligence Labs (Llama) @ Meta: I was the PM for one of the world's most important foundational models, focusing on the critical challenges of human alignment and safety.
- Validated at Scale: The principles in this course have been battle-tested and refined with over 2,200 product leaders—including 1,200 PMs & TPMs at Meta and over 1,000 others through my external workshops and classes.
- Instructor, Stanford Continuing Studies: I teach "Generative AI for Product Managers" &
- Building Generative and Agentic AI Products: A First-Principles Guide for Product Managers.
Building in the Front Lines: I have seen what works, what doesn't, and what's coming next in the AI space.
Previously at
