How Coinbase scaled AI to 1,000+ engineers | Chintan Turakhia

2026-03-02 · Source: How I AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Emerging Technologies & Innovation · Depth: Advanced, extended

Summary

Chintan Therakia, Senior Director of Engineering at Coinbase, discusses how his team successfully drove AI adoption and increased engineering velocity within a 1,000+ person organization. Facing the challenge of rewriting a product from a self-custody wallet to a social consumer app in 6-9 months, Therakia focused on using AI tools like Cursor to accelerate development. He emphasizes the importance of a hands-on leader with conviction, focusing on automating "toil" tasks like unit testing and linting, and creating a culture of sharing "wins." Key initiatives included a "PR speedrun" where 100 engineers generated 70 PRs in 15 minutes, and a company-wide speedrun with 800 engineers pushing 300-400 PRs in 30 minutes. He also details building an in-house agent, Cloudbot, to streamline the feedback-to-feature cycle, reducing PR review times by 10x from 150 hours to 15 hours.

Key takeaway

For Directors of AI/ML aiming to scale AI adoption in large engineering teams, prioritize a hands-on leadership approach to demonstrate tool efficacy and build conviction. Focus initial AI integration on automating repetitive, low-value tasks to generate quick wins and foster a culture of shared success. Consider implementing "PR speedruns" to rapidly showcase AI's impact on velocity and pressure-test your infrastructure, ultimately accelerating the feedback-to-feature cycle and reducing coordination overhead.

Key insights

Large engineering organizations can achieve significant AI adoption and velocity gains through hands-on leadership and strategic tool integration.

Principles

• Leaders must be hands-on to demonstrate AI tool value.
• Focus AI adoption on automating "toil" tasks for quick wins.
• Foster a culture of sharing AI successes to drive viral adoption.

Method

Implement "PR speedruns" to rapidly generate code changes and pressure-test infrastructure, using AI tools to automate mundane tasks and streamline the feedback-to-feature workflow.

In practice

• Use AI for unit test generation and linting.
• Build custom agents for specific security/workflow needs.
• Integrate AI tools into existing platforms like Slack and Linear.

Topics

• AI Adoption Strategies
• Engineering Productivity
• Developer AI Tools
• Feedback-to-Feature Automation
• AI Agent Development

Best for: Machine Learning Engineer, MLOps Engineer, Director of AI/ML

Related on AIssential

• How Coinbase scaled AI to 1,000+ engineers | Chintan Turakhia — Driving AI adoption in large engineering organizations requires hands-on leadership, automating toil, and data-driven user analysis.
• AI Coding Assistants Haven’t Sped up Delivery Because Coding Was Never the Bottleneck — AI coding tools shift bottlenecks from coding to specification and verification, demanding new team structures and workflows.
• This week on How I AI: Zapier's CEO shares his personal AI stack — Leaders must actively engage with AI adoption, providing tools and space for teams to build fluency and automate tasks.
• Quests, token leaderboards, and a skills marketplace: the elite AI adoption playbook | John Kim — Empowering non-technical teams with AI-building platforms accelerates internal innovation and fosters an AI-first company culture.
• EY: How to Scale AI at Speed Without Impacting Innovation — Effective AI architecture, not just model accuracy, drives speed, trust, and scalable human-AI orchestration.
• 🔮 Exponential View #561: Token scaling; frontier revenues; Spotify engineers; century bonds, T-cells vs Alzheimer’s & glass replaces silicon++ — Rapid AI adoption is transforming individual workflows and enterprise operations, necessitating a re-evaluation of AI's role.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by How I AI.