Hello

Sylvester was a collaborative research project I worked on during undergrad at the University of Detroit Mercy. It learned the language of Twitter through automatic annotation and classification, then interpreted tweets in real time to determine how people felt emotionally about any given subject, based on the current shape of online language rather than a fixed lexicon.

I built AI before you could use AI to build code. That predates the entire vibe-coding era. Working on Sylvester meant wrestling with the messy reality of unstructured language at scale: tokenization, drift, sarcasm, slang, ambiguity. The instincts I developed then are the foundation of what’s now called context engineering.

The last year and a half have been transformative, and the surface area of what’s possible keeps expanding. But I’ve watched every iteration of this from inside the work, not just as a user. If your team is wondering whether AI could be leveraged deeper into your product or workflow, I’m probably the right person to take that to the next step.