FinTech AI Research Intern: Transformer & GNN Prototyping

oneapp; United States; Hybrid; USD 34,000 - 55,000; Full time

Benefits offered by this job

Downtown NYC office

Hybrid work 3-4 days/week

Professional development activities

Team social outings

Job summary

oneapp is seeking a final-year graduate student for an applied AI research internship in New York City. You will explore transformer and graph-based models on fintech data, train and evaluate prototypes, and deliver a working demonstration with real customer impact.

The role embeds with product, engineering, and analytics teams to improve customer experience and business outcomes. Hybrid onsite 3-4 days weekly from June 21 to August 13, 2027.

Qualifications

• Final-year master's or PhD student graduating Fall 2027 or Spring 2028.
• Strong grounding in deep learning and model training.
• Experience implementing or fine-tuning models in PyTorch.
• At least one prior internship or project in ML/DS.

Responsibilities

• Design, train, and evaluate foundation models on real customer, transaction, and merchant datasets using transformer or graph architectures.
• Set up rigorous experiments, document results, and iterate based on quantitative findings.
• Translate research directions into a functioning prototype for product and engineering teams to build on.
• Collaborate with product, engineering, and analytics teams to align modeling with customer-facing outcomes.
• Communicate findings through summaries, presentations, and code.

Skills

Deep learning fundamentals

Experiment design

Communication

Education

Master's or PhD in engineering/CS

Tools

PyTorch

Graph neural networks

Job description

oneapp is seeking a final-year graduate student for an applied AI research internship in New York City. You will explore transformer and graph-based models on fintech data, train and evaluate prototypes, and deliver a working demonstration with real customer impact.

The role embeds with product, engineering, and analytics teams to improve customer experience and business outcomes. Hybrid onsite 3-4 days weekly from June 21 to August 13, 2027.