[ ABOUT THE ROLE ]
AgentOps is the discipline MLOps grew into. You will own the operational layer for live AI agents — monitoring, cost governance, drift detection, incident response, and the audit trails that keep enterprise clients confident. This is a high-stakes engineering role with direct client visibility.
[ RESPONSIBILITIES ]
- Build and operate real-time monitoring for multi-agent systems in production
- Implement token-cost dashboards and governance controls per client contract
- Design and maintain drift-detection pipelines with automated alerting
- Triage and resolve SLA incidents for L1–L3 support across client environments
- Write runbooks and post-mortems for every non-trivial incident
- Automate recurring operational tasks — reduce human-in-the-loop where safe
[ WHAT WE'RE LOOKING FOR ]
- 3+ years experience in DevOps, SRE, or MLOps roles
- Strong Linux, scripting (Bash/Python), and observability tooling (Prometheus, Grafana, or equivalent)
- Experience with containerised workloads (Docker, Kubernetes)
- Familiarity with cloud-native logging and tracing (GCP Cloud Logging, Datadog, or similar)
- Comfort with on-call rotations and structured incident management
- Clear, direct communication — you will work with both engineers and non-technical client stakeholders
[ NICE TO HAVE ]
- Prior experience running AI or ML systems in production
- Familiarity with LLM inference infrastructure (vLLM, TGI, or hosted APIs)
- Background building cost-allocation or chargeback systems
[ APPLY ]
Apply for this role