Location: London; Belgrade
Workplace: Hybrid
About the role
Perplexity's fraud prevention team protects our products, users, and infrastructure from fraud and abuse across the whole lifecycle: sign-up, authentication, payments, API usage, and product features. We build and operate a real-time decisioning platform together with the surrounding systems: data pipelines for detection and model training, integrations with product and payment systems, and internal tooling for investigation and response. As a software engineer on the team, you will own services end to end, from design to production operations, and work closely with fraud analysts, ML engineers, and product teams to make abuse expensive and legitimate usage frictionless.
Responsibilities
- Design, build, and operate high-throughput, low-latency backend services for fraud detection and prevention
- Build data pipelines and real-time feature computation that power detections and ML models
- Collect and process trust signals
- Integrate fraud decisions into product, auth, and payment flows across Perplexity
- Own reliability of production services: observability, alerting, safe rollouts, incident response
- Build internal tooling for investigation, enforcement, and auditability
- Partner with fraud analysts to turn new abuse patterns into production detections quickly
Qualifications
- 5+ years building and operating high-throughput, latency-sensitive backend systems in production
- Strong systems programming skills; Rust, Go, or C++ experience preferred (we write Rust)
- Hands-on experience with Kubernetes, Terraform and cloud infrastructure
- Solid data fundamentals: stream/batch computations, caching and storage trade-offs
- Pragmatic, ownership-driven approach: ship incrementally, measure, and make failures degrade safely
Nice to have
- Fraud prevention, anti-abuse, or payments-risk domain experience
- Experience serving ML models in production
- Client-side experience (web or mobile) for signal collection
Контакты работодателя доступны по кнопке «Откликнуться» после входа.
ГибридSan Francisco Bay Area, USA; Team members are expected to work from our San Francisco office three full days per week, with the remaining days worked remotely and are expected to be available during agreed working hours and to maintain sufficient overlap with the relevant team’s time zone
ГибридSan Francisco Bay Area, USA; Team members are expected to work from our San Francisco office three full days per week, with the remaining days worked remotely and are expected to be available during agreed working hours and to maintain sufficient overlap with the relevant team’s time zone