Joyce Chuyi Chen

I am an AI engineer and published foundation model researcher, journalist & poet. I currently lead voice at Inception, where I deploy, evaluate, post-train and align diffusion large language models for voice agents.

I studied mathematical & computational sciences and creative writing at Stanford, where I researched vision–language models, robot learning and participatory machine learning. I also revived and chaired the Editorial Board of The Stanford Daily.

Research & engineering

Inception — voice lead; post-training & deploying diffusion large language models. 2025–
Databricks · Mosaic AI — machine learning engineer; fine-tuning platform for enterprise LLMs. 2024
TEOChat — world's first vision–language model for temporal earth observation, Ermon Group. ICLR 2025
Other work: reward learning for robots in language space with ILIAD Group, participatory ML for representative content moderation with Stanford HCI Group, and computational tools to parse corpora of unstructured text for journalists at the Brown Institute for Media Innovation.

Prose & poetry

I write SF & Silicon Valley cultural criticism, book & movie reviews, and technical commentary on LLM development on my Substack.

My poetry has been broadcast on BBC national radio and shortlisted for the inaugural Sky Arts & Royal Society of Literature Writers Award.

“Senbazuru” — winner of the Young Romantics Prize, in The Keats-Shelley Review.
“Cuttlefishing off the Coast of Hong Kong” — a creative-critical response in emagazine and Forward Arts.
Poems — published by The Poetry Society.

Journalism

At The Stanford Daily, I revived and chaired the Editorial Board (Vol. 263 and 264), served on the Board of Directors, and was Managing Editor of Opinions.

Disturbing the universe — my senior column.
Fizz compromised its users’ privacy — an investigation into Stanford-founded social media startup’s security malpractice. Cited by TechCrunch.
Reuters — data journalist on the Enterprise Team, in collaboration with Big Local News.