Site Reliability Engineer

About Astera

Astera is a private foundation with a $2.5B endowment on a mission to steer science and technology toward an abundant future for all. Unlike traditional foundations, we operate like a high-velocity startup with unprecedented access to computational resources and complete freedom from funding pressures or profit motives. This allows us to focus on ambitious goals and attract incredibly creative scientists and engineers from leading academic institutions and from frontier AI labs.

Neuro-AI is our large-scale AI research program, pursuing a neuroscience-informed approach to engineering AGI. This is not yet-another-lab scaling LLMs in a hope of achieving general intelligence. We are integrating neuroscience, AI, and bioengineering to understand and digitally model the architecture of the human brain.

Position Summary

We are looking for a Site Reliability Engineer to own the digital infrastructure that powers our research.

This includes compute resources that we rent from third parties, container registries, and dashboards. The main objective is to make sharing these resources easy and efficient, ensuring the infrastructure is reliable and accessible to the right people.

This role spans a broad spectrum of activities:

• Compute Access: Ensure easy and efficient access to compute resources for our researchers.

• Resource Visibility: Provide clear visibility into resource utilization and cluster health.

• Auto-Scaling: Enable automatic scaling of compute resources based on demand.

• Access Management: Ensure the right people have access to the right resources.

• Reproducibility: Drive towards deterministic deployments and reproducible research environments.

• Process Automation: Automate operational processes where it makes sense to increase efficiency.

• Current stack: Ansible, Kubernetes, Docker, Tailscale, Python, Grafana, Prometheus, and Talos Linux. We're not religious about any of it.

Qualifications

• Ownership: You are comfortable being the person accountable when the cluster is unhealthy or capacity is tight.

• Systems Intuition: You understand how schedulers, containers, networking, storage, and hardware interact. You can reason about failure modes and design systems that degrade predictably.

• Operational Rigor: You value observability, reproducibility, and clear operational boundaries. You leave systems in a state that other engineers can understand, operate, and debug without you.

• Pragmatism: You can support experimental research workloads without forcing everything into a rigid "production" mold. You know when to stabilize and when to allow controlled chaos to speed up discovery.