Evals for taste: Hill-climbing a slide-generation agent
Built rubric-driven replayable eval system from real user projects giving quality/cost/latency/error/token signals in <6 hours per model change. Evolved into dev flywheel powered by real user dissatisfaction signals.
Details
City
Tokyo, JP
Date
11 June 2026
Time
13:00 – 13:45
Speaker(s)
Jonah Dueck
Anthropic
Agenda
08:00 – 09:00
Check-in and breakfast
09:30 – 09:45
Opening keynote
·
Angela Jiang
Cat Wu
Katelyn Lesse
morning sessions
10:00 – 10:30
Claude Code native company building: How GenerativeX rewired every function around AI
Founder stage
·
Rei Araki
10:00 – 10:45
10:45 – 11:15
From Claude prototype to production: How Myrealtrip builds and ships AI workflows
Founder stage
·
Wonjin Hur
11:00 – 11:45
11:30 – 12:00
12:00 – 12:45
11:45 – 13:30
Lunch
afternoon sessions
13:00 – 13:45
13:15 – 13:45
Same model, three different worlds: Japan, India, Australia
Builder stage
·
Jason Bigman
Rye Smith
Sumeet Doshi
Yuta Hayashi
14:00 – 14:30
Code less, query more: A co-founder's 30-day dive into Claude Code sub-agents
Founder stage
·
Kenta Yamamoto
How I built a legal platform for 280 million people at the Claude Code Hackathon
Builder stage
·
Ilham Firdausi Putra
14:00 – 14:45
14:45 – 15:15
15:00 – 15:45
15:30 – 16:00
Manager 11: How I finally shipped my dream football management game with Claude
Builder stage
·
Rye Smith
Evening
15:45 – 18:00
Closing reception
Anthropic's developer conference
Join us for a day of hands-on workshops, live demos of new capabilities and conversations with the teams behind Claude. Watch live from anywhere.