Remote GPU & ML Infrastructure Engineer

Svitla Systems, Inc.; Argentina; Híbrido; ARS 182.877.000 - 274.315.000; Jornada completa

Ventajas ofrecidas por este puesto de trabajo

Remote or office workspace

Technical webinars and meetups

Bonuses for talks and activities

Regular online activities

Descripción de la vacante

Svitla Systems, Inc. is seeking a GPU & ML Infrastructure Engineer for a full-time role in the USA. You will own end-to-end data generation for GPU systems, including benchmarking workloads, automated deployment, hardware telemetry, data quality checks, and dataset delivery.

You will work with NVIDIA data-center GPUs across generations and with edge platforms. The ideal candidate has strong Linux expertise, experience with LLM workloads on GPUs, and a track record of building reproducible GPU

Formación

• Strong Linux and GPU driver stack knowledge.
• Experience deploying LLM workloads on GPUs (training and inference).
• Experience collecting hardware telemetry programmatically (NVML/DCGM/IPMI/Redfish).
• Ability to work independently and overlap with client team until 11:00 a.m. PST.

Responsabilidades

• Port test procedures to new data-center GPUs and edge devices.
• Build and maintain GPU benchmark workloads incl. synthetic kernels and LLM workloads.
• Ensure telemetry collection is accurate across platforms and data sources.
• Investigate sampling issues, timestamp inconsistencies, and missing sensor data.
• Automate workload deployment, execution, data collection, and environment cleanup.
• Create automated checks for missing samples and clock mismatches.
• Deliver datasets with complete run metadata and documentation of hardware targets.

Conocimientos

Linux systems

GPU driver stacks

Python automation

Telemetry collection

Herramientas

NVML

DCGM

IPMI

Redfish

Descripción del empleo

Svitla Systems, Inc. is seeking a GPU & ML Infrastructure Engineer for a full-time role in the USA. You will own end-to-end data generation for GPU systems, including benchmarking workloads, automated deployment, hardware telemetry, data quality checks, and dataset delivery.

You will work with NVIDIA data-center GPUs across generations and with edge platforms. The ideal candidate has strong Linux expertise, experience with LLM workloads on GPUs, and a track record of building reproducible GPU