About the job
About Us
Greetings! We are Ravelin, a pioneering fraud detection firm leveraging state-of-the-art machine learning and network analysis technologies to tackle significant challenges. Our mission is to enhance the safety of online transactions, empowering our clients to serve their customers with confidence.
We also believe in enjoying the journey! Our team is welcoming and we take pride in our vibrant culture, embodying values such as empathy, ambition, unity, and integrity. We emphasize a healthy work/life balance and embrace a flat organizational structure. Join us to quickly learn about cutting-edge technologies while collaborating with some of the brightest and kindest individuals - check our Glassdoor reviews!
If this resonates with you, we are eager to connect! For further insights, visit our blog and discover how you can help us combat fraud and safeguard the world's leading online businesses.
The Team
You will be a vital part of the Detection team, dedicated to minimizing fraud rates and ensuring client satisfaction by continually training and deploying machine learning models. Our objective is to simplify model deployments to be as seamless and error-free as possible, adhering to Google's Best Practices for ML Engineering.
Our models are expertly trained to identify various types of fraud in real-time, utilizing diverse data sources and techniques. Our prediction pipelines operate under strict SLAs, ensuring every prediction is delivered within 300ms. When models underperform, it is the Detection team's responsibility to investigate the issues.
The Detection team is pivotal to Ravelin’s success, collaborating closely with the Data Engineering Team, who build the infrastructure, and the Intelligence & Investigations Team, who engage with our clients.
The Role
We are seeking a Data Scientist to train, deploy, debug, and evaluate our fraud detection models. The ideal candidate will be pragmatic, approachable, and equipped with insights gained from previous experiences.
Assessing fraud models can be challenging; often, labels may not be available for up to three months. You will need to exercise your judgment when exploring ambiguous fraud cases and validating the efficacy of the models themselves.
We strive to develop robust models capable of adapting their insights when exposed to new fraud tactics; our clients expect us to stay ahead of fraud trends. You will receive the tools, environment, and guidance necessary to create world-class fraud detection solutions.

