I’m a linguist and AI engineer working at the intersection of language and software. I build conversations, conversational AI systems, and the infrastructure around them: voice agents, synthetic data generation, evaluation methodology, annotation tooling.

I’m currently a Forward Deployed AI Engineer & Data Scientist at Five9, where I build production voice agents on AI Voice Agents and take on complex customer builds that help move the product forward. Before that, I spent four years at AWS supporting Amazon Transcribe and Contact Lens as a Language Engineer and then Language Data Scientist. I wore many hats while there, but my work at various points included agentic annotation systems, LLM evaluation methodology, and multilingual synthetic dialogue pipelines. I’ve co-authored work on LLM evaluation that has appeared at ACL, NAACL, and EMNLP.

My path here has run through linguistics: my PhD at UC Santa Cruz was an experimental investigation into the permeability of syntactic islands — mostly empirical and methodological work, extending an experimental paradigm and charting out an area where the standard picture of islands leaks (without settling the questions raised). That training persists, since turning fuzzy questions about linguistic behavior into measurable ones is still important to what I do.

Away from work, I maintain mkdnflow.nvim, a Neovim plugin for navigating markdown documents and markdown-based notes and wikis. I’m happiest when building — tools, knowledge, solutions to problems, or objects meant to be beautiful — and I’m gratified by continually improving those things.