About Castle

Castle is a Series A startup on a mission to create a safer online world by protecting platforms from fraud, abuse, and malicious activity. Trusted by companies like Canva, Atlassian, and Rockstar Games, we’re backed by Y Combinator, Index Ventures, and top-tier angels from Datadog, Stripe, and New Relic. Our fast-moving team is scaling globally, helping platforms stay secure while thriving in a rapidly evolving digital landscape.

The role

As a Machine Learning Engineer specializing in Bot & Fraud Detection, you will develop AI-driven fraud prevention systems that continuously adapt to evolving attack patterns. Your work will directly impact real-time bot mitigation, anomaly detection, and automated rule generation, ensuring Castle remains one step ahead of attackers.

You will work closely with our Head of Research, Antoine Vastel, a recognized expert in bot detection and browser fingerprinting, to integrate cutting-edge fraud detection models into Castle’s AI-driven defense engine. With direct access to rich behavioral data and a modern ML stack, you’ll be able to iterate quickly and deploy models that make an immediate impact.

In addition to traditional ML-based fraud detection, you will also leverage LLMs to classify traffic patterns, detect anomalies, and optimize application performance. You will experiment with the latest advancements in AI, ensuring Castle is at the forefront of modern fraud prevention technology.

What you’ll do

You will develop AI-powered fraud detection models, leveraging behavioral analysis, anomaly detection, and adversarial ML techniques to uncover evasive bot behavior. Our structured data pipelines and well-instrumented fraud signals ensure you’ll spend more time solving complex problems and less time wrangling data.

You will explore LLMs for traffic classification, using modern AI architectures to analyze session behavior, detect anomalies, and optimize fraud detection models. With seamless access to large-scale real-world fraud datasets, you'll have everything you need to push the boundaries of AI-driven fraud prevention.

What you bring

• 5+ years of experience in machine learning, fraud detection, or adversarial AI, preferably in security-focused environments.
• Strong background in anomaly detection, behavioral modeling, and adversarial ML techniques for detecting fraudulent activity and bot behavior.
• Experience developing and optimizing real-time ML pipelines for high-speed fraud detection.
• Hands-on experience deploying and fine-tuning LLMs, with a deep understanding of prompt engineering, fine-tuning, and large-scale inference.
• Deep understanding of supervised, unsupervised, and reinforcement learning techniques for adaptive fraud prevention.
• Familiarity with fraud signals, device intelligence, and behavioral analytics for bot detection and risk assessment.
• Experience with vector-based search, feature engineering, and scalable model inference for real-time fraud prevention.
• Ability to collaborate across Threat Research, Infrastructure, and JavaScript Engineering teams to develop end-to-end ML-powered fraud detection.

Benefits

• We pay US salaries globally. $153,000-$196,000 per year for this role.
• Flexible work hours. We prioritize outcomes over hours spent.
• Unlimited PTO. Take the time you need to recharge and maintain a healthy work-life balance.
• Paid parental leave. Supporting new parents during their transition.
• We’ll supply the computer and related gear you need to excel.