Meni Brief · Seattle, WA


Agents take actions for us, but the consequences are ours to bear.

The gap between those two is what I'm working on: running agents in worlds you control, to find out what they can actually be trusted to do, and when that trust should grow.
The rest of this site is the usual: a resumé if you need the credentials, rants when I have something to say, and work when there's something I can show.

who I am

Before moving to Seattle to be closer to the gorgeous mountains, especially the Cascades, I spent most of my life in Israel. I grew up in a small town that nobody has heard of, called Jerusalem, but after a short stint in Alaska I learned that what really matters is being close to nature. I love the mountains and thankfully don't have to suffer through hot summers anymore.

Growing up I always thought the technical stuff wasn't really for me, and that I wasn't very good at it. After spending a gap year studying philosophy and literature, I decided that it makes the most sense to study what I felt like I was worst at, so I went to the Technion (the self-proclaimed MIT of Israel). I was absolutely sure I would be a complete underdog, and didn't realize that I didn't suck at math until I somehow topped my class in the first year.

This led me on a path to discover I loved probability, statistics, data, and machine learning. I spent the next several years studying and researching these topics, and eventually found myself at Microsoft, where I worked on applied research. This was extremely lucky, since I happened to be perfectly positioned for the emerging field of LLMs (I worked with them before they were called LLMs, but back then they were just called Large Language Models). I was (and am) fascinated by the potential of these models, and how they could be used in real-world applications.

Seeing how fast things were moving in the world (especially compared to corporate tech environments), I decided I wanted to run off and build something myself. First came Simply-Useful, a task app for mobile-first teams that's live in the app stores. Then I went back to the itch left over from my Microsoft work: everyone is deploying agents, and almost nobody can say what theirs is actually trusted to do. Making AI actually useful means closing that gap, so that's what I'm building now. It's not ready to talk about yet, but it will be. I love building things, and I love seeing them used in the real world.

Away from the keyboard (or sometimes microphone button nowadays) I'm unsurprisingly into nature - hiking, skiing/snowboarding, surfing, etc. I'm also on my way to being a polyglot: English and Hebrew are native, and I get by in Spanish(+) and Levantine Arabic(-). Still determining what the fifth will be.

explore