AI Research Intern

Role overview This internship places a final-year graduate student on an applied AI research team investigating foundation models for financial services data. The intern will explore transformer-based and graph-based architectures, run controlled experiments, and deliver a working prototype that demonstrates a chosen approach. The work is embedded alongside product, engineering, and applied science teams, with the goal of improving customer experience, operational efficiency, and business outcomes.

Responsibilities - Design, train, and evaluate foundation models on real customer, transaction, and merchant datasets, working with transformer and/or graph neural network architectures. - Set up rigorous experiments, document results, and iterate quickly based on quantitative findings. - Translate promising research directions into a functioning prototype that product and engineering teams can build on. - Partner closely with product, engineering, and analytics colleagues to align modeling work with customer-facing outcomes. - Communicate findings clearly through written summaries, presentations, and code, contributing to shared team knowledge.

Requirements - Final-year master's or PhD student in engineering, computer science, or a closely related field, graduating in Fall 2027 or Spring 2028. - Solid grasp of deep learning fundamentals including backpropagation, attention mechanisms, embeddings, optimization, and training dynamics. - Hands-on experience implementing, training, or fine-tuning models in PyTorch or an equivalent framework, through research, coursework, or substantial personal projects. - At least one prior internship in research, machine learning, or software engineering, or equivalent project experience. - Availability for a full-time, 8-week program running June 21 through August 13, 2027, working onsite in a New York City office 3-4 days per week. - Ownership mentality, comfort with fast-paced environments, and strong organizational and communication skills.

Nice to have - Familiarity with transformer-based foundation model lifecycles, including pre-training, fine-tuning, inference, and architectural tradeoffs. - Exposure to graph-based learning approaches such as GNNs, message passing, or representation learning on graph-structured data. - Coursework or projects involving large language models or generative AI. - Research output in the form of publications, workshop papers, or open-source contributions.

Benefits and work setup - Hands-on experience shipping a research prototype with potential for production impact at a high-growth fintech. - Access to a downtown New York City office and structured professional development activities with team social outings. - Hybrid in-office requirement of 3-4 days per week for the duration of the 8-week program. - Visa sponsorship is not available for this position; candidates must be authorized to work without future sponsorship. - Standard interview process includes an initial screen, talent partner conversation, technical or hiring manager round, team interview, and executive conversation.