Praise for Roman Hauksson

“I also just have to say, your code is very exemplary in its organization and overall cleanliness. And so are your commits. I have a thing or two to learn from you!”

—research collaborator

“You're the most technically skilled member on our team.”

—former manager

“I really do mean it when I say you're more capable than you have a right to be at your age...”



  • University of Texas at Dallas

    Aug 2021 – Dec 2025 (Expected)

    • B.S. in Computer Science, National Merit Scholar

work experience

  • Capital One: Software Engineering Intern (McLean, Virginia)

    Jun 2023 – Aug 2023

    • Created front end for Chrome extension using React. Refactored build tool to increase preview speed by 300%.
    • Completed 4-week curriculum on full-stack technologies including Go, React, and Linux.
  • Joint Special Operations Command: Software Engineering Intern (Virtual)

    Jun 2022 – Aug 2022

    • Created Python interface for internal Node.js API.


  • Demo for "I"

    • Designed and created full-stack demo for a friend's startup, a trading platform for securitized income share agreements.
  • Manifold Markets Depth Chart Feature

    • Created React component to visualize the market depth of prediction markets.
    • Won 3rd place in Manifold Markets Hackathon and pushed it to production on a site with 800+ daily active users.

    • Fine-tuned GPT-3 on the lyrics of 96 MF DOOM songs to write lyrics in his style, then cloned his voice to edit those lyrics into a song.
  • Just Arrived – HackUTD VIII

    • Designed and coded front end for an app to help international students; won Nebula API Challenge.


  • ACM UTD Research Program: Director

    May 2022 – Dec 2022

    • Directed a semester-long program with a ~15% admission rate that gives undergraduates hands-on exposure to CS research.
    • Managed 4 teams of 6 students each and hosted the research presentation event for 50+ attendees.
  • ACM UTD Research Program: Team Lead

    Dec 2021 – Apr 2022

    • Fine-tuned GPT-3 on user data to explore the use of language models to generate targeted password guesses.
    • Achieved >3 times higher accuracy than baseline, won the research presentation contest versus 5 other teams, and amassed 61 stars on GitHub.
  • ACM UTD Research Program: Participant

    Sep 2021 – Nov 2021

    • Using Tensorflow, trained an LSTM model on Tweets from different subgroups on Twitter with a team of 4.
    • Won the research presentation contest versus 5 other teams.


  • Program on AI and Reasoning (Cambridge, UK)

    Aug 2023

    • Learned about AI and cognition at a 10-day camp for mathematically talented students (2.5% acceptance rate).
  • We'll Pwn You With Your Wattpad Profile (New York, New York)

    Jul 2022

    • Delivered talk on password cracking at the Hackers on Planet Earth conference to an audience of 20+ hackers.
  • TexSAW 2022 Capture The Flag Winner (Dallas, Texas)

    Apr 2022

    • Won a cybersecurity CTF competition with a partner against 10+ teams.