This is the first entry in a new research notebook. I want a place to record ideas, engineering lessons, and questions that arise while working across AI systems and computer architecture.

My interests include AI accelerators, LLM serving systems, speculative decoding, and agent systems. These topics often meet at the same technical boundaries: computation, data movement, scheduling, and the abstractions between them.

Future posts may cover reading notes, project designs, and experiments that can be shared publicly. I will distinguish what a source reports from what an experiment shows and what remains an open question.

This post is the starting point. Project overviews are available under Projects, and research directions under Research.