AI Research Intern

s for Spring or Summer 2027. Research interns conduct research relating to frontier AI detection research or related topics, with the goal of publishing a paper in a major conference venue. Examples of past internship projects at Pangram include EditLens (https://arxiv.org/pdf/2510.03154) and Scaling Laws for Wild AI Text (https://arxiv.org/abs/2609.40295).

What You’ll Work On

Interns are assigned a mentor, and choose a research topic or area prior to the internship. Current active research areas within Pangram include:

• Adversarial Robustness: how do detectors perform against humanizers and adversarial paraphrasing?
• Interpretability: can we build more explainable, interpretable AI detection?
• Mixed Human/AI Authorship: how can we quantify the different facets of human-AI coauthorship?
• New Architectures: how can we optimize Pangram’s architecture for efficiency or even higher levels of accuracy?
• Multimodal Detection: how can we improve AI detection for images, video, and audio?
• Effects of AI-Generated Data on LLM Training: what is the effect of AI-generated data on foundation model training?
• Agentic AI detection: how can we harness the power of agents to give Pangram more context about authorship?
• Topics of your choice: propose an original topic relevant to AI detection

Requirements

• Currently pursuing an M.S. or Ph.D. in Computer Science/Artificial Intelligence. Exceptional undergraduates entering their senior year are welcome to apply.
• Prior research experience that has resulted in a published paper in a top AI venue.
• Fluency in deep learning fundamentals and experience training neural networks.
• Excellent programming skills in Python and a deep learning framework like Pytorch.
• Experience working with an AI agent for research like Claude Code or Codex.

Compensation

• $8,000-$11,000 per month
• Relocation stipend available
• In-office lunch
• Top interns may receive return offers for full-time roles