Detection Automation Engineer, Bot & Fraud Detection
Castle
WorldwideremotePosted 2 months ago
Skill Required
Fraud-Detection-EngineeringBot-DetectionTrust-And-Safety-EngineeringSecurity-EngineeringDetection-Systems-EngineeringFraud-Detection-EngineerSenior-Fraud-Detection-EngineerAnti-Fraud-EngineerThreat-Detection-EngineerautomationDesign SystemEngineeringbuildingdesignRustandMachine LearningFulltime
Key highlights
- US‑level salaries paid globally.
- Remote‑friendly for Europe‑based candidates.
- Must have experience building large‑scale bot or fraud detection engines.
- Ability to think adversarially and design clear, maintainable systems.
- Focus on outcomes rather than hours worked.
Role overview
Castle is a small, profitable team building a real-time trust layer for modern platforms. We are seeking an experienced engineer who has built and operated real‑world bot or fraud detection engines at scale, understands the adversarial nature of fraud, and can design systems that are simple, observable, hard to game, and resilient to change.
Responsibilities
- Design detection systems that are simple enough to reason about, observable in production, hard to game, and able to evolve without collapsing under their own weight.
- Build and maintain a real‑time trust layer for modern platforms that survives browser releases, new automation frameworks, fingerprint manipulation techniques, and traffic growth.
- Understand attacker behavior and design systems that make entire classes of abuse harder, not just the latest variant.
- Manage latency constraints and business tradeoffs while ensuring legitimate users are not blocked.
- Evaluate when machine learning genuinely improves outcomes versus when deterministic or statistical approaches are cleaner and more robust.
- Think adversarially, prioritize clarity, and create systems that survive contact with reality.
Requirements
- Proven experience working on real‑world bot or fraud detection engines at scale.
- Experience blocking traffic at high volume.
- Experience dealing with false positives that impact real users.
- Deep understanding of latency constraints and business tradeoffs in detection systems.
- Ability to discern when machine learning adds value versus when deterministic or statistical methods are preferable.
- Adversarial mindset and a focus on building clear, maintainable detection systems.
Nice to have
- History of encountering a detection model that accidentally blocked a major website or app and learning from that incident.
- Experience designing systems that survive major browser releases and evolving automation frameworks.
- Familiarity with fingerprint manipulation techniques and how to mitigate them.
- Background in creating observable, production‑ready monitoring for detection pipelines.
Benefits
- US‑level salaries paid globally.
- Remote‑friendly work environment, especially for candidates in Europe.
- Focus on outcomes over hours worked.
Additional details
- Fraud detection is adversarial systems engineering; it’s a cat‑and‑mouse game.
- while (true) {attackers.adapt();signals.decay();browsers.change();models.drift();rules.stopWorking();}
- Common responses—adding more rules, hiring more analysts, tuning thresholds, stacking additional models—lead to fragile, opaque systems with overlapping rules and increasing false positives.
- Castle prefers detection systems that remain simple, observable, and robust without devolving into rule spaghetti.
- This slower‑starting approach compounds over time, yielding more sustainable protection.
- Originally posted on Himalayas.