Building the most AI-pilled engineering team in the world

Fiona Fung leads the Claude Code and Cowork teams at Anthropic and previously ran Visual Studio, Facebook Marketplace, and Meta's AR/VR and Instagram work. In this conversation she discusses how software engineering and engineering management are being reshaped by AI, including new verification practices, hiring profiles, async workflows, and how to keep teams healthy and product-focused.

01Key learnings

• Manage with AI by running recurring Claude sessions over repos, Slack, and metrics to summarize shipped work and turn it into coaching conversations rather than just tracking output.
• Automate daily leadership rituals such as scanning feedback channels with scheduled routines that surface themes and draft PRs for review, reducing manual toil.
• Give Claude explicit frameworks such as checked-in specs and skills describing what good looks like, so automated code review can validate changes against them.
• Hire two profiles: creative builders with product sense who own products end to end, and deep systems experts for areas where verification still needs real expertise.
• Measure impact through outcomes rather than lines of code or tool usage; use shared quality thresholds such as bad (irrecoverable) versus sad (recoverable pain points) to focus attention.
• Use the product yourself daily as a leader, and when you are new to a team, spend time as an IC learning the codebase before taking on management duties.
• Watch for latent demand where users stretch a product in unexpected ways, and be willing to productize those emerging use cases.
• Periodically challenge existing processes and kill those that no longer serve their purpose; Fiona shifted from six-month roadmaps to lightweight monthly just-in-time planning.

02Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.