Kevin Li
marketer / founder / judoka
// Now
- On a career break. I left jo in August 2026 after three years building a personal AI agent, back when it was a bet and not a race. No next thing yet, and that's the point.
- Training for black belt in judo. Currently brown belt, next competition March 2027. My tokui-waza is tai otoshi.
- Living in Alameda, CA with my 10-year-old daughter Kery and family.
- Playing Wild Rift and Path of Exile 2 when I should be sleeping.
// Building
Building towards a judo black belt. After 15 years of building companies, I'm building the one thing that doesn't ship. Training most days: judo at the RTC, early-morning BJJ, strength in between. Startups taught me to move fast. Judo is teaching me to move well.
Not building software right now, but still thinking about how humans and AI should work together. That question isn't going anywhere.
// Journey
I'm a two-time YC founder with 15+ years in growth marketing. It started at Kabam. I joined as an intern in 2008, back when the company was still called Watercooler, and stayed through its pivot into social gaming.
My first startup was Second Wave Games (2011). It sold to World Golf Tour (later acquired by Topgolf), where I stayed on as Head of Mobile.
Then product marketing at MessageMe (2014), the mobile messaging app Yahoo acquired. My boss Arjun Sethi was a real mentor to me in those early startup years.
At Yahoo I was a Product Manager for Growth & Emerging Products (2014–2016). I grew Yahoo Mail into a top-4 app overall through growth experimentation. That's where the growth playbook came from: test everything, keep what moves the needle, never ship without an experiment behind it.
Then I co-founded Farmstead (YC S16), an AI-powered "ghost grocer" that delivered groceries in under 60 minutes from small-format micro-warehouse hubs.
We used machine learning for demand forecasting and kept food waste under 5% (supermarkets average around 35%). It started as a Nextdoor post in Mountain View offering weekly milk, eggs, yogurt, and bread. Two hundred replies in two days. Then COVID hit, orders went 10x, and we scaled to a $10M+ run rate across five markets before winding down in 2023.
Then I co-founded jo, a personal AI agent, where I was CMO. We were early, before the personal-agent race had a name. I left in August 2026.