Harden the production agent dispatch platform by implementing reliable routing, circuit breakers, retries, surfaced errors, caching, and event streaming. Own orchestration and deployment decisions, introduce infrastructure as code and per-session isolation, and establish platform-wide standards for testing, CI, observability, and service scaffolding.
Optimize customer training and inference workloads on committed GPU clusters by identifying bottlenecks, establishing hardware-specific performance baselines, and working with customers or operating teams on tuning. Encode optimization expertise into agent-executable workflows, while also owning internal model serving improvements and advising on GPU allocation.
Own the reasoning and orchestration architecture for AI agents that diagnose infrastructure issues, recommend or take actions, and explain their decisions. Build agent-behavior evaluation and autonomy SLOs, establish evidence-based governance for expanding agent permissions, and encode operational runbooks into agent-executable reasoning.
Design and build a unified, agent-queryable model of fleet assets, topology, configuration, and current state, and create a pipeline that captures operator decisions. Own the telemetry access layer and scale path, while coordinating with telemetry-owning teams on signals, granularity, retention, and raw event access.
Build and improve the platform’s reliability and scalability through caching, event streaming, resilient agent dispatch, and service-level observability. Contribute to orchestration and infrastructure-as-code migrations, and strengthen deployment practices with canary releases, traffic shifting, and reliable rollback.
The Technical Success Manager serves as the primary operational point of contact for managed-services customers, coordinating between internal teams and the client. They are responsible for managing customer relationships, overseeing service performance, tracking lifecycle requirements, and facilitating recurring business reviews.
The Director of Data Center & Site Operations leads global physical infrastructure delivery, overseeing both initial build-outs and steady-state operational activities. This role manages deployment schedules, vendor performance, and operational readiness to ensure high availability and scalability across multiple regions.
The SRE owns reliability, observability, and incident response for the GPUaaS platform. Key duties include defining SLOs, building the observability stack, and leading major incident resolution.
Build and operate the multi-tenant orchestration and scheduling layer to transform raw GPU infrastructure into a cloud service. Design customer-facing APIs, CLIs, and automation for node provisioning and image management.
The NOC Engineer manages 24/7 monitoring and first-response for GPUaaS infrastructure to protect customer SLAs. Responsibilities include triaging alerts, executing runbooks, and coordinating with on-call specialists during incidents.
