AI Engineer- Responsible AI

Join a pioneering AI safety startup to design and deploy robust security systems against LLM jailbreaks, adversarial attacks, and agentic vulnerabilities. You will own projects from research to production, building scalable safety infrastructure and evaluation pipelines.

• Advance AI safety through attack/defense strategies for LLMs. Build distributed safety evaluation pipelines handling millions of requests. Design ETL pipelines for petabyte-scale safety data. Evaluate human-AI interaction patterns to uncover behavioral vulnerabilities. Develop agentic AI security workflows. Deploy safety models using containerized microservices and CI/CD. Create automated evaluation protocols for model safety and adversarial resilience.

• Master's degree in CS/EE/ML/Security (PhD preferred). 2+ years industry experience in applied ML/AI research or engineering. Publication record in AI safety, NLP robustness, or adversarial ML. Strong Python and PyTorch/JAX skills. Experience with LLM jailbreak attacks/defense, agentic AI safety, adversarial ML, or human-AI vulnerabilities. Containerization (Docker, Kubernetes) and cloud platforms.

• Experience with adversarial prompt engineering, jailbreak detection. Multi-agent architectures or robust defense strategies for LLMs. Large-scale data processing (Spark, Flink, Kafka). MLOps expertise (Triton, TensorRT, vLLM, W&B, MLflow). Infrastructure as Code (Terraform, Pulumi). Distributed computing (Ray, Dask). Observability stacks (Prometheus, Grafana, DataDog). First-author publications or strong open-source contributions.