The role
Patronus AI contracts experienced SREs, platform and observability engineers for about 10 hours a week. You'll work incidents live with researchers, and your solutions become evaluation and training data.
What you'll do
- Work through incident and infrastructure scenarios live with a researcher
- Explain your reasoning, including fixes you considered and rejected
- Review drafted scenarios and flag what's unrealistic or underspecified
- Write acceptance criteria, including tempting wrong moves under pressure
- Write short post-session notes and act on reviewer feedback
What they're looking for
Must-haves
- 3+ years running production systems, with on-call in the last two years
- Hands-on with Kubernetes, Prometheus and Grafana (or Datadog, New Relic)
- Experience with a log stack (Loki, ELK, Splunk), Terraform, Helm and CI/CD
- Comfortable in Python, Go or Bash and reading unfamiliar code
- Clear spoken English and reliable availability for scheduled live sessions
Nice to have
- Experience writing runbooks, post-mortems or SLO definitions
- Familiarity with how AI coding agents fail on operations work
- Experience creating incident training material
Skills and platforms
- Kubernetes
- Prometheus
- Grafana
- Helm
- Terraform
- CI/CD
- Site Reliability Engineering
- Incident Response
- Python
- Go
- Bash
- Observability
- LogQL
- PromQL
- Infrastructure as Code
- Debugging
The company
Patronus AI
Technology, Information and Internet
Patronus AI is a model behavior research lab developing training data and simulated environments for frontier AI. Our work spans evals and digital world models for agents operating in software.
Summarised by Jobs That Travel from Patronus AI's listing.
