Evals are the product spec.
If you can't write down what good looks like as test cases, you don't have a product yet, just a demo.
Nate Richardson·PM × Growth engProduct × Growth engineering
I build AI products— and the growth engines behind them.
Product manager who ships. I turn LLMs, agents and data into products people adopt and pipelines that compound — then I measure whether it worked.
Ask my agent· answers from my work, with sources/
- activation from [AI onboarding]
- +[XX]%
- activation from [AI onboarding]
- LLM agents running in production
- [N]
- LLM agents running in production
- pipeline from [GTM automation]
- $[X]M
- pipeline from [GTM automation]
- growth experiments shipped
- [NN]
- growth experiments shipped
02Point of view
AI made building cheap. Distribution is the moat.
When anyone can ship the feature, the advantage moves to whoever owns the loop that gets it to users.
PMs should be able to ship.
Not to replace engineers, but because a prototype ends a debate faster than a doc does.
03Experience
- [2023]–Now
[Product Manager, Growth] · [Company]
[One line on the product, team and what you owned.]
- [Shipped X, which moved Y by Z.]
- [Built an internal LLM tool that saved N hours/week.]
- [2020]–[2023]
[Title] · [Company]
[One line on scope.]
- [Outcome with a number.]
04Contact
Building something with AI? Let's talk.
Product management or growth/GTM engineering roles at teams building with AI. [Stage, size or domain preferences.]