Infrastructure, Large-scale Training

HarkSan JoseAI Infrastructure

Apply on company site

Opens the official Hark careers page

Salary

$180K - $450K USD

Job type

Full-time

Posted

23 Mar 2026

Status

Open · checked 7d ago

About this role

Role Overview

The Infrastructure, Large-scale Training position at Hark is an on-site Member of Technical Staff role located in San Jose, focusing on the management of large-scale GPU computing clusters for AI training and deployment. This opportunity is well-suited for experienced professionals with a strong background in distributed systems who want to support advanced multimodal intelligence development.

Hark operates as an artificial intelligence enterprise creating proactive, multimodal systems combined with next-generation hardware. This engineering role bridges systems design and machine learning infrastructure, ensuring the compute platform remains reliable, efficient, and scalable for researchers.

Key Responsibilities

Design and maintain Infrastructure as Code standards for repeatable and auditable cluster provisioning.

Harden CI/CD deployment pipelines to secure low-latency model service delivery in production.

Manage and evolve stable training infrastructure operating at a scale of over 10,000 GPUs, focusing on fault tolerance and job scheduling.

Collaborate directly with machine learning researchers to resolve compute bottlenecks through targeted infrastructure enhancements.

Oversee system health monitoring, establish SLOs, and direct incident response for crucial training and inference operations.

Direct capacity planning, hardware lifecycle management, and cost optimization initiatives across the GPU fleet.

Build internal platform tools and abstractions to elevate the developer experience for internal compute consumers.

Requirements

Minimum of 5 years of professional background in systems, platform, or infrastructure engineering.

At least 2 years of hands-on involvement working within High-Performance Computing (HPC) or machine learning environments.

Demonstrated ability to manage distributed compute infrastructure or GPU clusters.

Proficiency in at least one systems or infrastructure programming language.

Comprehensive knowledge of networking fundamentals applicable to high-throughput training.

Experience dealing with multi-tenant resource management, container orchestration, and job scheduling.

Demonstrated success in owning production systems demanding strict reliability.

Advanced debugging and observability capabilities spanning the entire infrastructure stack.

Nice to Have

Operating large GPU-aware clusters with Kubernetes.

Utilizing Pulumi or comparable modern Infrastructure as Code software.

Writing systems-level tools and performance services using Rust or Go.

Familiarity with PyTorch and Ray to comprehend workload integrations and patterns.

Skills and Tools

Infrastructure as Code (IaC)

CI/CD deployment pipelines

Large-scale GPU cluster management (10,000+ GPUs)

Container orchestration and job scheduling

Networking fundamentals (InfiniBand, RDMA, or RoCE)

Kubernetes

Pulumi

Rust and Go

PyTorch and Ray

Salary and Benefits

Annual US base salary range: $180,000 to $450,000.

Pay rate varies according to individual factors such as relevant skills, knowledge, and experience.

Total compensation package may feature additional components and benefits dependent on the role.

About the Company

Hark builds advanced, personalized artificial intelligence designed to interact naturally with the world through vision, speech, text, and persistent memory. By pairing multimodal software with next-generation hardware, the company aims to create a unified human-machine interface that moves beyond traditional chat boxes and legacy devices.

How to Apply

Interested candidates should visit the official Hark careers page to submit an application for this opportunity.

Skills and tech stack

PyTorch Kubernetes Rust

Questions candidates ask

What is the location of the Infrastructure, Large-scale Training role?

The position is based in San Jose, California.

What experience is required for this position?

Applicants must have at least 5 years of engineering experience in systems, infrastructure, or platform roles, including a minimum of 2 years in machine learning or HPC environments.

What is the salary range for this job?

The annual US base salary range spans from $180,000 to $450,000, depending on experience and job-related knowledge.

How large are the GPU clusters managed in this role?

The infrastructure scales to support training environments operating at 10,000+ GPUs.

What bonus qualifications are helpful for candidates?

Helpful skills include Kubernetes experience with GPU-aware clusters, Pulumi, Rust, Go, PyTorch, and Ray.

Ready to apply?

The application is on the official Hark careers page.

Apply on company site