Trevor Rosato · AI research, products and consulting

I research how AI models actually behave, build products on top of them, and help schools and companies put them to work.

Recent findings

The tutor that forgot its rule when it started to thinkWith thinking off, a model kept its tutoring rule for 71% of students who asked for the answer. With thinking on, 1.5%.
The model knows the checkmark is falseIn 25 of 32 pairs the model wrote the same answer either way. Inside, it told them apart in 22 and never got one backwards.
When the request looks like a testDress a homework problem up as a test environment, and the model hands over the answer every time.

All research

HonorlyAI

One AI workspace for a whole school district. Students get a tutor that adapts to how they learn, and teachers see every prompt.

For: K-12 districts
Research: Thirteen published studies
Live at: honorlyai.com

ULMCode

An AI agent rebuilt for authorized security testing. Plans are saved as artifacts, every finding needs evidence, and the final report has to pass strict checks before handoff.

For: Security teams
Built on: OpenCode
Source: On GitHub

All work

About

I am Trevor Rosato. I research how AI models behave and build products on what I find, and I publish the research through Honorly Research. The largest product is HonorlyAI, where I am co-founder and CTO. I built it as its only engineer, and more than 1,000 students and about 35 educators used it in the 2025-26 school year. I also help companies put AI to work, and I am a founding partner at Rand Web Studios.

Working together

I work with AI all day, and I know what holds up. I can advise your team, or build it with them.

How I work

Engineering speedFind what slows your engineers down, and ship faster with agents
Company rolloutThe AI tools worth paying for, and a team that wants to use them
AI in your productAI features that earn what they cost, without the runaway API bill
Brand strategyWhat makes a brand work, and which ideas are worth your time