Site Reliability Engineer
Beam is an AI-native serverless cloud platform for GPU inference, task queues, and secure sandboxes.
Funding history
About Beam
Smartshare, Inc. operates Beam, a developer platform for running AI and ML workloads—including inference endpoints, agents, task queues, and sandboxes—on CPUs and GPUs without managing underlying infrastructure.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will own end-to-end compute fleet health by building metrics pipelines, alerting, and a unified GPU health view. You will automate failure detection and triage, design GPU qualification processes, and maintain firmware telemetry and large-scale log collection. You will also investigate production incidents through post-mortem.
Requirements
- Knowledge of hardware failure modes at firmware and silicon levels
- Knowledge of AI tooling
- Ability to debug production issues from triage through post-mortem
- Knowledge of developer tools, cloud-native technologies, and open-source software
Responsibilities
- Build metrics pipelines, alerting, and a unified GPU health view
- Automate compute failure detection and triage
- Design GPU burn-in, performance baselining, and qualification processes
- Own firmware-level telemetry and large-scale log collection
- Debug production issues from triage through post-mortem
Benefits
- Meaningful equity
- Health, dental, and vision benefits with 90% coverage for employees and 50% for dependents
- Opportunities to participate in cloud-native community events
- Fitness stipend
- Learning budget
