Staff+ Software Engineer, Account Compromise at Anthropic

on-site · full-time · Visa sponsorship

Responsibilities:
- Set technical direction and own architecture for account compromise detection, response, and remediation across Claude and the Developer Platform.
- Scope and lead complex, multi-month engineering projects from ambiguous start to reliable production systems.
- Build and evolve near-real-time detection systems for account takeover, credential abuse, and compromised API keys.
- Design automated response flows that cut off attacker access while minimising disruption to legitimate users.
- Lead end-to-end investigations for significant compromise incidents and convert findings into durable automated defenses.
- Threat model attacker behaviors, prioritise work against likely attacker adaptations, and align with Security, Product, and Policy teams.

Requirements:
- 10+ years building production detection, security, or anti-fraud systems.
- Strong proficiency in Python and SQL; experience with big-data tooling (BigQuery, Spark, dbt, Airflow) and ML-based signals preferred.
- Deep understanding of authentication/identity systems (OAuth, SSO, MFA) and API key security.
- Demonstrated incident leadership, adversarial problem-solving, and architecture-level decision making.
- Track record of delivering reliable systems under adversarial pressure and balancing false positives vs. user impact.
- Ability to operate autonomously and drive cross-functional alignment.