7 months ago
Responsibilities
- Train end-to-end models for fault prediction, network state modeling, and autonomous repair.
- Build multimodal models over structured networking data and use function-calling to make network decisions.
- Evaluate model performance on real-world hardware and virtualized environments.
- Ship models into production networks and iterate using real-world operational data.
- Collaborate with firmware and application engineering teams.
Requirements
- Experience building, training, and scaling neural networks from the ground up.
- Systems-oriented thinking beyond benchmark performance.
- Interest in modeling the physical world and collaborating across hardware and software.
- Interest in helping establish the technical direction of a new applied research organization.
Benefits
- Access to thousands of H100 GPUs for model development.
- Direct collaboration with the founders and executive team.
- Ground-floor opportunity to shape the roadmap, modeling approaches, team culture, and long-term vision.
- End-to-end ownership from model development through production deployment in real-world networks.
- Opportunity to work with proprietary networking data and full-stack control over hardware, firmware, operating systems, and operations.
Categories
About Meter
Meter builds and operates managed internet infrastructure for businesses, combining enterprise networking hardware, cloud software, and deployment/monitoring services. The company designs and runs wired and wireless networks as a turnkey service for offices and multi-site enterprises. Founded in 2015 and headquartered in San Francisco, it is privately held and counts organizations such as Bridgewater, Lyft, and Reddit among its customers.