Hello! I'm a Research Engineer in Safety & Alignment at Character AI. I train and test models to increase safety and interpretability. I'm a loud advocate of integrating cross-discipline oversight into data labeling, reward modeling, and evaluation design.

Previously, I recieved my PhD in Computer Science at the University of Southern California, co-advised by Professors Kristina Lerman and Fred Morstatter. My work studied community-grounded approaches to dialect distribution drift tasks with subjective answers. I've also interned with Amazon, designed ISI’s STEM summer program, and obtained my B.S. in Computer Science at UC Santa Cruz.

Let's connect! Feel free to take a look at my CV and reach out to me at rdorn[at]usc[dot]edu.

Recent News

Oct 2026 I will be joining members of c.ai's technical team in joining the Walt Disney Corporation.
June 2026 This month, I start as a safety and alignment research engineer with Character AI.
Feb 2026 Defended my PhD Dissertation, Community-Grounded Measurement and Mitigation of Dialect Bias in Subjective Tasks!!
Nov 2025 Had an incredible time at the EvalEval symposium, where I presented ongoing work about the transferability of dialect bias mitigation techniques between subjective datasets.
May 2025 Started my internship with Amazon Science, where I'm building a human-in-the-loop evaluation agent for newly generated content.
Apr 2025 This month, I'm honored to be a speaker at Microsoft's GLEAM lecture series about LGBTQ+ in Cybersecurity research, and Mila's Fairness and Transparency in AI (FATE) Seminar.
Dec 2024 The Center for Democracy and Technology uses QueerReclaimLex to help highlight intersectional disparities in content moderation. (Someone pinch me I'm dreaming)
Nov 2024 Wow, our OATH-Frames paper received an outstanding paper award at EMNLP 2024! I've learned so much from this interdisciplinary project. Kudos to my hardworking collaborators!