Senior Specialist, Data Scientist

BNY•Pittsburgh, PA

2d•$71,000 - $123,000•Onsite

About The Position

We're seeking a future team member to join our Payments Risk Services team as a Data Scientist / Data Engineer. In this role you'll learn our payments data end-to-end and build machine learning models that detect and prevent fraudulent payments. This role is located in Pittsburgh, PA, and is well-suited to a recent college graduate eager to apply data science to a high-impact, real-world problem.

Requirements

Bachelor's degree in a STEM field (Computer Science, Data Science, Statistics, Mathematics, Engineering, or related), or equivalent experience.
Foundational knowledge of machine learning and statistics, and hands-on coding in Python with libraries such as scikit-learn, pandas, and a deep-learning framework (PyTorch/TensorFlow).
Familiarity with SQL and working with large datasets; exposure to distributed/cloud data tools (e.g., Spark, cloud data platforms) is a plus.
Curiosity about fraud detection, anomaly detection, or risk analytics — a demonstrated approach to understanding a problem domain and translating it into data-driven solutions.
Interest in or exposure to explainable AI (XAI) and responsible/ethical AI practices.
Strong problem-solving and communication skills, with the ability to explain technical results to non-technical audiences.

Nice To Haves

Internship, academic project, or coursework experience in ML, data engineering, or the financial services industry is a plus.

Responsibilities

Learn the payments data landscape — Explore, profile, and understand transaction, customer, and channel data across payment rails (wire, ACH, RTP/instant) to build the foundation for detection models.
Build fraud-detection ML models — Develop, train, and validate supervised and unsupervised models (classification, anomaly detection, graph/network analysis) that flag fraudulent payments in batch and near-real-time.
Develop a fraud typology-driven approach — Understand the major categories of payments fraud — account takeover, authorized push payment (APP)/scams, synthetic identity, business email compromise, money mule/laundering patterns — and map each to detection signals and modeling strategies for how to detect and address them.
Engineer features and data pipelines — Design and maintain reliable feature pipelines (behavioral, velocity, device, network, and aggregate features) that feed models, partnering with data engineering to move from prototype to production.
Leverage AI to strengthen detection — Identify opportunities to apply modern AI techniques (e.g., deep learning, embeddings, LLMs for unstructured signals, foundation/graph models) to improve fraud coverage and reduce false positives.
Build explainable AI (XAI) — Apply model-interpretability methods (SHAP, LIME, counterfactuals, reason codes) so fraud analysts, model risk, and regulators can understand why a payment was flagged.
Measure and communicate impact — Track model performance (precision/recall, false-positive rate, fraud dollars prevented), and clearly present findings and recommendations to both technical and business stakeholders.
Own projects from inception to delivery — Partner with fraud SMEs, product, and engineering to take detection ideas from hypothesis through deployment and monitoring.
Stay current — Follow fraud trends, emerging attack patterns, and advances in ML/AI and responsible-AI practices relevant to the banking industry.
Grow across the data science domains — Build depth in model science, feature science, and insight science, strengthening core skills in programming, math & statistics, distributed computing, and communicating complex results.