Prime Intellect
Member of Technical Staff - Training Platform
$150K–$300K
San Francisco +1 · Remote · Staff · Full-time · Posted Jul 8, 2026
- Location
- San Francisco
Remote - Visa sponsorship
- Full visa sponsorship and relocation support.
About the job
Build and operate a hosted platform for launching and managing fine-tuning runs on GPU clusters, spanning Kubernetes infrastructure, backend services, and developer-facing tools.
What you'll do
- Develop and operate Kubernetes-based training infrastructure, including orchestration, scheduling, autoscaling, storage, and observability.
- Build platform backend services, APIs, and user interfaces for submitting jobs and monitoring training runs.
- Integrate training, inference, and environment systems, and productize new models, training methods, and capabilities.
What they ask for
- Strong working knowledge of modern AI models, fine-tuning, inference engines, GPU hardware, and distributed training fundamentals.
- Strong Kubernetes operations experience, including Helm and production cluster debugging.
- Strong Python backend development and experience building control-plane agents that interact with Kubernetes APIs.
- Experience with modern frontend development and REST or tRPC API design.
Nice to have
- GCP experience; GCP is preferred.
- Experience building developer tools, dashboards, and live-monitoring UIs.
Pay
$150K–$300K USD base salary
Tools
Kubernetes, Helm, KEDA, FastAPI, Python, TypeScript, React, Prometheus
AI focus
Build infrastructure and product capabilities for post-training AI models, including LoRA and full fine-tuning, RL, inference, and support for new model architectures and algorithms.
Model-written summary · the original posting is authoritative