Systems & Research Engineer – Applied AI (Fully Remote, Australia)

About the Role

Join a leading US applied AI company specializing in advanced AI production systems within the freight and global supply chain sector. Operate remotely from anywhere in Australia—Canberra, Sydney, Melbourne, Brisbane, Perth, Byron Bay—or globally, without relocation or strict office hours. The company embraces asynchronous work practices, valuing engineering quality, research rigor, and sound decision-making over synchronous status.

Key Responsibilities

This role demands engineering excellence at the crossroads of AI systems, research, inference, model serving, performance engineering, distributed systems, and evaluation. You will systematically investigate the real-world behavior of production AI systems to identify bottlenecks involving GPU, CPU, memory, network, architecture, or concurrency. Your work will include hypothesizing, benchmarking, testing alternatives, and making data-driven engineering decisions that impact production. Examples of tasks include benchmarking speech-to-text systems leading to hybrid models outperforming commercial vendors in quality and cost-effectiveness.

Expectations & Approach

Beyond generic AI engineering, the position requires the application of the scientific method to production AI environments. Candidates must be able to articulate experiments encompassing baseline establishment, hypothesis formation, benchmarking, controlling for bias, data analysis, and the subsequent design or architecture changes based on findings.

Technical Scope

• Profiling AI system components to identify performance bottlenecks
• Benchmarking serving frameworks like vLLM and SGLang
• Evaluating inference throughput and latency
• Comparing open-source and proprietary models
• Optimizing workloads balancing quality, cost, and concurrency
• Developing infrastructure for evaluation and understanding model behavior with live traffic
• Engineering distributed systems integrating AI models distinctly in the processing loop
• Translating research insights into scalable production architectures
• Challenging assumptions with experimental evidence

Who Should Apply

Ideal candidates may currently hold titles such as Research Engineer, ML Systems Engineer, AI Infrastructure Engineer, Inference Engineer, Performance Engineer, ML Platform Engineer, or Systems Engineer. Experience within sophisticated machine learning infrastructure contexts similar to those in major tech companies or advanced AI startups is highly valued. Critical is the ability to communicate detailed experimental outcomes and engineering impacts clearly.

What This Role Excludes

This role does not encompass DevOps, frontend development, conventional full-stack work, Web3 infrastructure, or purely distributed systems engineering without meaningful AI/model involvement.

Technical Philosophy

The team extensively utilizes coding agents to elevate engineering efficiency, focusing on design, architecture, validation, and decision-making over manual coding. Comfort and proficiency with these modern tools will be evaluated during interview stages.

Work Arrangement

This is a bona fide fully remote position, supporting workers globally with no mandatory office attendance. While the company maintains a San Francisco office, use is optional. Strong written communication, ownership, and asynchronous collaboration are essential.

Compensation

The total remuneration can reach approximately A$420,000 plus equity options. This corresponds roughly to a US$220,000–300,000 salary range with US$300,000 as the maximum. This pay scale offers Australian residents compensation commensurate with top-tier US AI engineering roles.

The Hiring Standard

The hiring bar is intentionally stringent, seeking candidates who surpass the calibre of the existing engineering team. Exceptional achievements in AI systems, inference, model serving, evaluation, or performance engineering are prerequisites. Applications should emphasize the experimental process and decisions rather than just project outcomes.