Drimva

ONE Finance

AI Research Intern

United States (Remote) · Remote · Internship · Posted Sep 21, 2026

Experience
At least one previous internship in research, machine learning, or software engineering, or substantial personal or academic projects.
Education
Final year of a master’s or PhD program in Engineering or a related field; expected graduation in Fall 2027 or Spring 2028.
Visa sponsorship
OnePay will not sponsor applicants who need employment sponsored visas for this position. Future sponsorship will not be considered.

About the job

Researches and prototypes transformer- and graph-based foundation models using customer, transaction, and merchant data, evaluating experiments and developing a working prototype.

What you'll do

  • Research, design, and train foundation models on OnePay data.
  • Run and rigorously evaluate experiments, turning promising results into a working prototype.
  • Collaborate with Product, Engineering, and Analytics teams on customer experience and business outcomes.

What they ask for

  • Final-year master’s or PhD student in Engineering or a related field, expected to graduate in Fall 2027 or Spring 2028.
  • Strong understanding of deep learning fundamentals and experience implementing, training, or fine-tuning models; comfortable with PyTorch or an equivalent framework.
  • At least one prior internship in research, machine learning, or software engineering, or substantial personal or academic projects.
  • Available for the full-time, 8-week program June 21–August 13, 2027, and to work in the New York City office at least 3–4 days per week.

Nice to have

  • Familiarity with transformer foundation models or graph-based learning and graph foundation models.
  • Coursework or projects involving Large Language Models or Generative AI; research publications, workshop papers, or open-source contributions.

Pay

$32 per hour

Tools

PyTorch, transformer-based architectures, graph-based architectures, GNNs, Large Language Models, Generative AI

AI focus

Research and prototype foundation models, including transformer- and graph-based models, applied to customer, transaction, and merchant data.

Model-written summary · the original posting is authoritative

First seen Oct 29, 2026 · Last verified Nov 5, 2026 (UTC)