Laurence
RL for performance marketing
Advertising today is like trading in the 1980s pits: archaic, manual, and iterative guesswork. Humans update bids and keywords using noisy, delayed data with fixed rules that both break at scale and ignore stochasticity.
We've built quantitative systems that advertise the way modern hedge-funds trade markets: continuous decision-making under stochastic outcomes and explicit profit constraints.
We use customers' existing ad copy and use reinforcement learning and train our own LLMs to run ads on autopilot, harvesting profits when we’re confident and borrowing signals from similar keywords when data is sparse.
Starting with brands on Amazon, we are automating millions in live ad spend, increasing gross sales while decreasing advertising cost of sales by 40%.
Active Founders
Founder
Built the Data Science Agent at Google. Grew Facebook new user retention by 18% while tuning Facebook's friending model. Third engineer at health-tech startup. Math+CS & AI research at Cornell. Ex-national team and semi-pro soccer goalie.
Washed dishes and waited tables to pay for college. Financially independent at 17. National team & semi-pro soccer goalie. Taught myself to code at 18. Chemistry research at 15.
CEO
Built the Data Science Agent at Google. Grew Facebook new user retention by 18% while tuning Facebook's friending model. Third engineer at health-tech startup. Math+CS & AI research at Cornell. Ex-national team and semi-pro soccer goalie.
Washed dishes and waited tables to pay for college. Financially independent at 17. National team & semi-pro soccer goalie. Taught myself to code at 18. Chemistry research at 15.
CEO
Company Launches
Laurence - Automated Amazon Advertising
See original launch postHey YC! 👋
We’re Matt and Leo building Laurence: automated Amazon pay-per-click (PPC) performance marketing using reinforcement learning and transformer models.
TL;DR
- Most brands lose money advertising on Amazon after margin
- Agencies outsource simplistic copy-paste playbooks to offshore workers
- Existing software are rudimentary rules engines that break with any bad variance
- We drive performance using reinforcement learning and training our own conversion-rate-aware transformer models
- Our quant system is automating $10 million in live ad spend, increasing advertising profits by 15-20% while growing total revenue.
Team
Matt
- built the Data Science Agent at Google
- tuned the Facebook friend suggestion model at Meta
- was the third engineer at a health-tech startup
Leo
- made $250k as a high-stakes poker pro at 17
- founded a crypto investing startup at 18
- was a researcher at Jump Trading
We started off building a horizontal data science agent for enterprise, and from our hundreds of conversations with anyone who had a problem with data, we learned how desperately brands on Amazon needed PPC automation that works.
Ask
If you know any CPG brands or SMBs selling on Amazon, we would love an introduction :)
- Matt & Leo W26
