Teams at Together AI

Operate and scale multi-petabyte distributed storage for AI and HPC workloads. Design high-performance parallel filesystems, object stores, caching architectures, and Kubernetes-native storage operators. Optimize data paths, throughput, multi-tenancy, cost, and reliability across thousands of GPU nodes. Integrate technologies such as Vast, Weka, Ceph, and Lustre; develop automation and tooling in Go and Python; and contribute to storage infrastructure, observability, benchmarking, and open-source projects.

Build production software for automated GPU and inference cluster provisioning. Design declarative APIs, control planes, state machines, reconciliation loops, durable workflows, and self-healing systems that manage hardware from discovery through deployment, operation, repair, and decommissioning. Partner with inference and ML platform teams to model cluster topology and scheduling requirements. Own reliability, testing, CI/CD, deployment, and production operations for the platform.

Operate and scale Together AI's production infrastructure: run GPU-enabled Kubernetes clusters, automate with Ansible and Terraform, build monitoring/observability, respond to incidents (on-call), debug production issues, design deployment/upgrade processes, and plan infrastructure growth for high availability and scalability.