Principal Technologist – AI Compute

This is a fully remote position, open to applicants in United States.

📋 Description

• Take ownership of the technical architecture for intricate, high-stakes customer projects involving AI/HPC GPU and accelerator compute cluster design, workload performance, and the compute-to-network interface.

• Collaborate with Sales and Solutions Architects from initial discovery to deal closure.

• Lead the design and execution of proof-of-concept projects.

• Establish criteria for proof-of-concept success and develop benchmarking plans for GPU clusters focused on training/inference throughput and scaling efficiency.

• Present proof-of-concept outcomes to customers with precision and clarity.

• Interact with ML infrastructure leads, compute architects, and executive stakeholders at hyperscalers, NeoClouds, service providers, and large enterprises.

• Gather insights from the field regarding GPU/accelerator platform trends and workload behavior.

• Convert field insights into actionable requirements for Product Management and Engineering.

• Create and distribute architecture playbooks, reference designs, and best practices for AI compute infrastructure.

• Mentor and guide Solutions Architects and Solutions Engineers.

• Represent DriveNets at industry conferences and events.

• Write white papers and technical blogs.

• Enhance DriveNets' external technical reputation in the AI compute infrastructure sector.

⛳️ Requirements

• Over 12 years of experience in designing and architecting data center compute infrastructure.

• A minimum of 3 years dedicated to AI/HPC GPU or accelerator platforms in hyperscale settings.

• Comprehensive hands-on experience with GPU/accelerator hardware and system design, particularly with NVIDIA/AMD.

• Proficient in compute cluster orchestration utilizing Kubernetes and Slurm.

• Familiarity with distributed training/inference frameworks.

• Demonstrated experience in senior pre-sales, solutions architecture, or system architecture roles.

• Direct involvement in influencing large, complex deals with VP and C-level executives.

• Knowledge of GPU virtualization and partitioning, including MIG/vGPU.

• Experience in managing containerized ML workloads and provisioning bare-metal GPUs.

• Skilled in scripting and automation with Python, APIs, and JSON.

• Outstanding communication and presentation abilities.

• Willingness to travel domestically and internationally up to 10% of the time.

• Familiar with PyTorch, TensorFlow, and JAX.

• Experience with NCCL/RCCL tuning and GPU cluster benchmarking, such as MLPerf.

• Understanding of collective communication dynamics at scale.

• Practical knowledge of scale-up interconnects, including NVLink and UALink.

• Familiarity with GPU resource scheduling and orchestration using Slurm, Kubernetes, and Ray.

• Experience in GPU observability and telemetry with tools like DCGM, Prometheus, and Grafana.

• Knowledge of data center operations fundamentals, including power, cooling, and rack design.

• Holds a BS, MS, PhD, or equivalent experience in Electrical/Computer Engineering, Computer Science, Physics, or another Engineering discipline.

🏝️ Benefits

• Opportunity to significantly impact multi-million-dollar customer projects.

• A creative, collaborative, and growth-focused work environment.

• Approximately 10% travel, both domestic and international.

• Commitment to equal opportunity employment.

• Advantageous NVIDIA/AMD platform certifications or equivalent qualifications.