About Us

Pastel is an enterprise AI company building the intelligence layer for modern banking. We transform compliance, risk, and transaction data into real-time fraud detection, AML, credit decisioning, and natural language executive insights.

Founded at Stanford, we partner with financial institutions to move beyond static dashboards and manual reviews. Our platform helps banks detect risk faster, reduce compliance friction, and unlock growth opportunities hidden inside their own data.

We are building category-defining infrastructure at the intersection of AI, regulation, and financial services. If you want to work on hard, meaningful problems that directly contribute to financial inclusion, you’ll feel at home here.

Role Summary

We are seeking a Data Scientist to drive the development and deployment of machine learning models that power Pastel’s compliance and risk intelligence products.

This role requires both technical depth and strategic oversight. You will design systems that detect fraud, flag anomalies, and learn in real time as millions of transactions flow through our platform.

You will lead model experimentation, optimization, and deployment while working closely with our engineering, data, and product teams to turn AI research into production-ready impact.

Key Responsibilities

Machine Learning & Modeling

Design, build, and train advanced machine learning models for fraud detection, risk scoring, anomaly detection, and customer behavior prediction.
Perform extensive feature engineering, feature selection, and representation learning to improve model performance and interpretability.
Develop time-series models, anomaly detection pipelines, and predictive systems tailored to financial transaction data.
Leverage feature stores to ensure consistent, reusable, and production-ready feature pipelines across training and inference.

LLM & Agentic AI Systems

Develop and integrate LLM-powered pipelines for document intelligence, transaction reasoning, alert summarization, and investigative workflows.
Design agentic AI systems capable of multi-step reasoning, tool use, and automated decision support within compliance and risk contexts.