Student, Haverford College, PA
August 2011 — May 2015
Bachelors of Science in Computer Science & Bachelor of Arts in Linguistics. My undergraduate thesis was on the use of the Viterbi algorithm for real-time sign language recognition
About Me
Hi I am Maggie Perkoff 
,
I am a Postdoctoral Researcher at the University of Pennsylvania working on the U-GAIN Reading Project. I have a joint PhD in Computer Science and Cognitive Science from the University of Colorado at Boulder. My advisors were Dr. James Martin and Dr. Clayton Lewis. I am interested in leveraging generative AI to increase educational equity for learners everywhere. Additionally, I am passionate about response AI usage both inside and outside of the classroom setting. Previously, I have worked with the NSF Institute for Student AI Teaming, the American Institutes for Research, Nookly AI, and Amazon Alexa.
Software Engineer I, Amazon Alexa, Seattle, WA
July 2015 — September 2017
Software Engineer for the Amazon Alexa Conversational AI Team. I worked to expand our contextual carryover features to regions outside of the United States and support new languages on the device including French, German, and Japanese.
Software Engineer II, Amazon Alexa, Seattle, WA
September 2017 — July 2020
Software Engineer for the Amazon Alexa Conversational AI Team. I worked to expand our contextual carryover features to regions outside of the United States and support new languages on the device including French, German, and Japanese.
PhD Student, University of Colorado at Boulder
August 2020 — May 2025
I received a joint PhD in Computer Science & Cognitive Science at the University of Colorado Boulder. My two excellent academic advisors are Dr. James Martin and Dr. Clayton Lewis. I was a Research Assistant with the NSF Institute for Student AI Teaming as well as the TalkMoves project. My dissertation focused on the use of generative AI to supportive collaborative learning in middle school classrooms.
NAEP Doctoral Intern in AI, American Institutes for Research
May 2024 — September 2024
I worked with NAEP Mathematics data to assess the viability of using machine learning models to score constructed response NAEP mathematics items for grades 4 and 8.
Machine Learning Intern, Nookly AI
October 2024 — January 2025
At Nookly, I designed and implemented a RAG system to support a parenting assistant chatbot. The chatbot could provide basic support to parents as well as refer Nookly users to appropriate story building tools to support SEL skill building.
Postdoctoral Researcher, University of Pennsylvania
August 2025 — present
Presently, I am a postdoc at the University of Pennsylvania working on the [U-GAIN Reading Project](https://ugain-reading.org/). I'm investigating how to leverage generative AI to adapt reading instruction to individual learner's needs.