Senior AI Safety Engineer - Production Guardrails & Abuse Detection

ElevenLabs; Polska; Remote; PLN 551,000 - 866,000; Full time

Benefits offered by this job

Discretionary stipend for personal and

Annual team offsite

Learning & development stipend

Co-working stipend

Job summary

ElevenLabs is seeking an experienced AI Safety Engineer to design and deploy scalable backend systems for automated moderation and guardrails across multimodal spaces. You will collaborate with ML and full-stack engineers to bring production-ready safety infrastructure to life.

The role emphasizes end-to-end technical execution, from architecture to deployment and monitoring, with a focus on detecting abuse of free tiers. Remote-friendly with global teams and multiple office options.

Qualifications

• 6+ years of backend software engineering experience.
• Production-grade systems, scalable architectures for real-time and batch workloads.
• Experience deploying ML/AI systems in production.
• Strong Python (async) and backend frameworks.
• Cloud platforms (AWS/GCP) and containerized deployments.
• Observability with Prometheus and Grafana.

Responsibilities

• Design and build scalable backend infrastructure for abuse detection and production ML models.
• Architect APIs and data pipelines for real-time and batch moderation.
• Implement monitoring and observability with SLIs and SLOs.
• Collaborate with ML engineers to translate research models into production systems.
• Drive safety roadmap and guardrails at scale.

Skills

Backend engineering

Python

Distributed systems

APIs

Observability

ML/AI systems in prod

Tools

AWS/GCP

Docker/K8s

CI/CD

Prometheus

Grafana

Kafka

Job description

ElevenLabs is seeking an experienced AI Safety Engineer to design and deploy scalable backend systems for automated moderation and guardrails across multimodal spaces. You will collaborate with ML and full-stack engineers to bring production-ready safety infrastructure to life.

The role emphasizes end-to-end technical execution, from architecture to deployment and monitoring, with a focus on detecting abuse of free tiers. Remote-friendly with global teams and multiple office options.