LogoFresh Jobs Ahead
STN Incorporated
STN Incorporated
11 – 50 Zaposleni
Kibernetska varnostPodjetjeStrojna oprema
STN Incorporated provides managed IT and cloud infrastructure for enterprises and teams building business-critical software, AI systems, and robotics. Its services include private CPU clouds, GPU One AI infrastructure, secure networking and storage, and ongoing cloud operations supported by 24/7 human support and high-uptime service-level agreements. The company also handles cybersecurity operations and incident response, compliance and risk management, backup and recovery, and technology procurement and lifecycle management. STN serves SaaS companies, AI builders and model developers, robotics firms, and regulated organizations including healthcare providers, with infrastructure designed to support audit readiness and frameworks such as SOC 2 Type II and HIPAA.
LinkedIn
Delovna mesta v podjetju

Senior Inference Workload Optimization Lead

Lead GPU performance tuning for customer AI workloads on STN’s autonomous data center platform. Turn optimization expertise into agent-executable fixes and oversee internal model serving.

USZdružene države AmerikeNa daljavoPolni delovni časSeniorZahtevani Jezik: angleščinaUmetna inteligenca

Opis

  • Analyze customer training and inference workloads to locate performance bottlenecks across data loading, collective communication, memory bandwidth, kernels, batching, and parallelism.
  • Set performance baselines for each hardware generation, beginning with current-generation NVIDIA systems.
  • Translate optimization expertise into fixes that agents can apply across customer clusters.
  • Tune workloads with committed-capacity customers, directly or through their operating teams.
  • As a secondary responsibility, manage internal model serving, including cost- and latency-aware routing and model-version migrations.
  • Advise on allocating GPUs between internal inference and customer workloads.

Zahteve

  • Bring deep GPU performance engineering expertise, including profiling, memory and bandwidth behavior, and kernel-level understanding.
  • Have hands-on knowledge of serving stacks such as vLLM, SGLang, or TensorRT-LLM, including batching, KV cache management, and quantization trade-offs.
  • Have experience with distributed training and inference at scale, including tensor, pipeline, and data parallelism.
  • Have experience tuning collective communication, such as NCCL.
  • Be fluent with profiling tools such as Nsight, DCGM, and framework-level profilers.
  • Be able to explain performance findings to customers and encode solutions for an agent.
  • Nice to have: CUDA kernel development, large-scale open-weight model deployment, benchmark design, and previous customer-facing technical experience.

Sorodna delovna mesta

CDW

AI Infrastructure Consulting Engineer, Compute and Storage

CDW
Sponzor vizuma H1BIzobraževanjeSvetovanjeZdravstveno varstvo
USZdružene države Amerike
77.725 $ – 108.265 $ / leto
Na daljavoPolni delovni časSrednja ravenSeniorZahtevani Jezik: angleščinaUmetna inteligenca
KubernetesLinuxNFS

Design and validate GPU compute and high-performance storage solutions for enterprise AI and machine learning workloads. Help CDW customers implement infrastructure for AI training, fine-tuning, and inference.

Odpri