C++ Systems Engineer — GPU Virtualization & Performance

Uncover; San Francisco, Northern (CA, KY); Hybrid; USD 180,000 - 260,000; Full time

Benefits offered by this job

Competitive salary

Equity

Lunch daily

401(k)

Health insurance

Job summary

CompanyThunder Compute is building a general-purpose GPU virtualization layer, enabling GPUs to be pooled and allocated like networked resources. You will own core C++ systems, focusing on latency, reliability, and production readiness, with research into new allocation and oversubscription techniques.

You will work on profiling, debugging, and extending support across CUDA apps and hardware, collaborating with customers and cross-functional teams to deliver scalable GPU infrastructure in

Qualifications

• Strong C++ programming skills with focus on low-latency, high-performance systems.
• Deep understanding of operating systems and distributed systems concepts.
• Experience building production-grade C++ services for GPU or networked systems.

Responsibilities

• Profile and reduce latency for remote CUDA operations.
• Build high-performance networking and data-transfer paths.
• Debug failures across customer processes, runtime, network, and remote GPUs.
• Improve support for process forking, signals, multithreading, and dynamic linking.
• Design systems for GPU allocation, scheduling, and observability.
• Research and deploy GPU virtualization and oversubscription techniques.
• Expand compatibility across CUDA applications and GPU architectures.

Skills

Modern C++

OS concepts

Networking

Linux

Debugging

Production systems

Tools

CUDA

Dynamic linking

Performance profiling

Job description

CompanyThunder Compute is building a general-purpose GPU virtualization layer, enabling GPUs to be pooled and allocated like networked resources. You will own core C++ systems, focusing on latency, reliability, and production readiness, with research into new allocation and oversubscription techniques.

You will work on profiling, debugging, and extending support across CUDA apps and hardware, collaborating with customers and cross-functional teams to deliver scalable GPU infrastructure in