GPU Systems Engineer (AI)

Location: Austin, TX, United States; Chicago, Illinois, United States; London, United Kingdom; New York, NY, United States

Base salary: 150,000–300,000 USD per year, plus discretionary performance-based bonuses and benefits.

Company

hirify.global applies a scientific approach to financial trading and operates a large-scale computing environment for algorithmic trading research and development.

What you will do

• Design, build, and optimize distributed GPU compute clusters for HPC and AI workloads.
• Identify and resolve performance bottlenecks across compute, storage, and networking layers.
• Profile, benchmark, and fine-tune GPU-based workloads with research and development teams.
• Automate system deployment, monitoring, and troubleshooting across thousands of nodes.
• Own critical infrastructure projects from concept through implementation and ongoing support.
• Test and deploy hardware and software, partnering with vendors to resolve complex issues.

Requirements

• 5+ years of large-scale Linux systems engineering experience in HPC, AI, or distributed infrastructure.
• Extensive experience with Linux installation, performance tuning, and troubleshooting.
• Expertise in distributed GPU workload troubleshooting and GPU optimization.
• Proficiency in Python scripting and automation frameworks.
• Experience with NVIDIA technologies including NCCL, GPUDirect RDMA, and NVLink.
• Familiarity with configuration management tools such as Salt, Ansible, Puppet, or Chef, and the ability to diagnose hardware, operating system, and network issues.

Nice to have

• CUDA or C/C++ experience.

Culture & Benefits

• Collaborative environment spanning research, engineering, and infrastructure teams.
• Work on high-impact automation and computing challenges in algorithmic trading.
• Culture focused on openness, transparency, diverse expertise, and togetherness.
• Competitive benefits package and eligibility for discretionary performance-based bonuses.
• AI tools are prohibited during interviews or assessments unless explicitly authorized.