Berkeley ––:––
Beijing ––:––

which end is thinking right now 此刻我在哪一端思考

Keyi Li 李 柯 毅

Where product intuition meets agent infrastructure — I close the gap between intelligence and experience. 产品直觉智能体基础设施的交汇处——我弥合智能与体验之间的鸿沟。

keyi_li@berkeley.edu github.com/laokaoya

both, always两者兼具,始终如此

move cursor to explore移动鼠标探索

Fields explored探索的领域

Perpetual explorer持续探索者

00 / 04

Select a domain to see every story. 点击一个方向,探索所有故事。

click any node 点击任意节点

Experience Library经历档案

Click any node to expand点击任意节点展开详情

The person behind the résumé简历背后的人

Spin-½ particles never point in one direction. Neither do I — and I've stopped pretending that's a problem. 自旋½的粒子从不固定方向。我也是——已经不再假装这是个问题了。

My Xiaohongshu handle is 半自旋 (spin-half) — named after particles that can't commit to a single axis. Feels accurate. I started at PKU in geochemistry (China's most notorious "career sinkhole"), escaped into economics, then AI. Somewhere in between I spent a birthday on an Italian glacier staring at 4.5 billion years of exposed rock and thought: geological timescales are genuinely therapeutic. Best thinking happens between 2 and 4am. This is a feature. 小红书名叫"半自旋"——量子力学里那种不能固定轴向的粒子。感觉很贴切。本科在北大读了地球化学(传说中的天坑之王),后来转经济,又转了AI。中间有一年在意大利阿尔卑斯的冰川上过生日,对着45亿年的岩层发了很久的呆——这种尺度感对焦虑很有帮助。我的最佳思维时间是凌晨两点到四点。这是feature。

Geology fieldwork · 2am product spirals · Go (enthusiastically mediocre) · Mountains that reset timescales · Cameras pointed at things nobody notices 地质野外 · 凌晨两点脑洞 · 围棋(热情高于水平) · 能重置时间感的山 · 拍没人注意的东西

UC Berkeley
M.S. Statistics统计学硕士
2025 — present至今
Peking University
B.Sc. Chemistry & Economics · Minor CS · Top 1%化学 & 经济学理学学士 · 辅修 CS · 专业前 1%
2021 — 2025 · survived the sinkhole· 从天坑爬出来了

Honors荣誉

PKU Honor Graduate '25 · National Math Competition 2nd Prize · Internet+ Entrepreneurship 2nd Prize · PKU Outstanding Student '23 & '24 北京大学优秀毕业生 2025 · 全国大学生数学竞赛二等奖 · "互联网+"创业大赛二等奖 · 北大三好学生 2023 & 2024