Software Engineer, Bot & Fraud Detection Platform at Castle

About Castle

Castle is building real-time infrastructure to protect SaaS platforms from fraud, abuse, and malicious automation. Our product stops bots, detects account abuse, and helps teams fight sophisticated attacks without slowing down real users.

We’re trusted by companies like Canva, Atlassian, and Rockstar Games, and backed by Y Combinator, Index Ventures, and top-tier angels from Stripe, Datadog, and New Relic.

The role

Castle’s detection platform is only as strong as the system behind it. You’ll architect and own the pipelines, rules engines, and fingerprint intelligence systems that power our core product.

This is a hands-on role with high leverage. You won’t be implementing dashboards or frontend features; you’ll be building the substrate that enables us to detect and adapt to real-world attacks at scale.

You’ll work directly with our Head of Research (Antoine Vastel) to automate detection logic, integrate adversarial ML and anomaly detection, and drive the evolution of our internal data architecture.

This is one of the most foundational roles in the company.

What you’ll do

• Design and optimize Castle’s detection infrastructure for speed, resilience, and adaptivity

• Build real-time pipelines that ingest and process behavior, fingerprint, and fraud signals

• Create automation for rule deployment and fingerprint updates with millisecond latency

• Develop a dynamic detection layer that can reconfigure in response to novel attacks

• Apply vector-based search and event-driven architectures to classify and cluster threat behavior

• Enable seamless developer access to detection primitives—without exposing detection secrets

• Collaborate tightly across detection, ML, product, and infrastructure to keep attackers out and users safe

What you bring

• 5+ years of experience in detection engineering, bot mitigation, or security data pipelines

• Deep understanding of stream processing and event-driven systems (Kafka, Flink, etc.)

• Experience building low-latency, real-time systems for detection or automation

• Strong grasp of fingerprinting, reputation scoring, and adversarial attack patterns

• Experience with high-performance DBs like ClickHouse, Redis, or similar

• Familiarity with ML tools (e.g., anomaly detection, clustering, vector search)

• A platform builder’s mindset—you want to enable others, not just build features