Spent the morning with @RepAdamSmith and a strong group on defense + AI. The defense-AI conversation is finally catching up to where America's adversaries already are.
We're hiring across the board. Rather than list every role (link below), a few things worth knowing about the company:
1/ We believe AGI can go well, and that requires more defensive technology
built for the enterprises that critical infrastructure runs on.
2/ We're shipping investigative agents at scale, deployed inside some of the world's largest companies.
3/ The work sits at the intersection of frontier research and applied AI.
4/ Fully in-person, opinionated, highly technical team in San Francisco.
If any of this resonates, come build with us.
In this episode of Founder Firesides, YC Managing Partner Jared Friedman talks to Karine Mellata (@karine_exe), co-founder of Variance (@trustvariance), who is coming out of stealth and announcing their $21 million Series A.
Variance builds purpose-built AI agents for risk and compliance — automating fraud detection, content review, and identity verification for Fortune 500 companies and platforms like GoFundMe.
They discuss why Variance built in the shadows for three years, detecting state-sponsored fraud rings, and the accident that nearly ended the company.
00:49 – The AI That Keeps the Internet Safe
01:28 – Why They Stayed Secret for 3 Years
02:26 – You’ve Used This Without Knowing It
02:57 – How GoFundMe Stops Scams
03:59 – How Scammers Use Big News Events
05:50 – Checking IDs and Businesses Online
07:44 – How the AI Agents Work
09:28 – The Hardest Problem: Bad Data
12:07 – Why This Only Works Now
14:22 – Catching Organized Fraud Groups
16:26 – Tiny Team, Huge Output
20:18 – How They Met at Apple
22:24 – Getting Their First Customer
24:57 – Recovering from Getting Hit by a Truck
29:36 – Sticking to One Big Idea
In 2025, $4.4 trillion was spent on illegal activity through the global financial system.
$1.1 trillion went to drug trafficking.
$500 billion went to human trafficking.
$16 billion went to terrorist financing.
We raised $21.5 million to stop that.
alphabet came after our name, so we have a new name. meet variance!
we took the opportunity to rebrand and double down on what we actually do: deep autonomous investigations for adversarial environments
Reminder that our AI speakeasy happy hour is tomorrow! We've got a great lineup of speakers - come meet fellow builders and hang out with us.
📍 Mission District, SF
📆 Wednesday, Feb 12th
Reminder that our AI speakeasy happy hour is tomorrow! We've got a great lineup of speakers - come meet fellow builders and hang out with us.
📍 Mission District, SF
📆 Wednesday, Feb 12th
We're co-hosting a special speakeasy happy hour with @trustintrinsic on Wednesday, February 12th. You'll hear from our stellar guest speakers and meet other AI builders and tinkerers. Come join us in the Mission District in SF!
Okta Ventures portfolio company @TrustIntrinsic is on a mission to create a safer internet by democratizing safety tooling.
We sat down with their Co-Founder and CEO @KarineMellata to learn more about how they work with Okta 📝 https://t.co/erhLmTrU7n
Congrats to the team @trustintrinsic on the $3.1M seed!
Founded by a team that worked on solving Trust & Safety issues for Apple's 1 billion users, Intrinsic (YC W23) enables companies to automate these processes, from observability to enforcement. https://t.co/7gCmMTWf3v
Welcome to YC W23, @karinemellata, @mlin_disco, and @trustintrinsic!
Intrinsic allows companies to build and maintain high-quality content communities, monitor brand safety, and keep their users safe.
Learn more: https://t.co/CAdVERP5iW