Performance Engineer (C++, Python, Rust)

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We help people start and grow durable freelance careers teaching AI by consolidating specialized projects, building a consistent portfolio, and making it easy to apply and earn.

• We focus on the human side of AI: annotating, evaluating, and improving training data that real models learn from.
• OpenTrain hires contractors to create, review, and improve training tasks and evaluations used across cutting-edge AI systems.

Why AI training work matters

AI training (data labeling and human evaluation) is how modern models learn to behave: people design tasks, produce solutions, and rate model outputs so systems improve safely and effectively.

This line of work is remote, flexible, and accessible — contributors shape model behavior and see measurable impact on real systems.

• Remote and flexible work with opportunities to build technical expertise in a fast-growing industry.
• Work directly on training data and evaluation that influence frontier AI systems and inference infrastructure.

Role overview

OpenTrain is recruiting a Performance Engineer to write and evaluate performance-engineering tasks and solutions used in AI training and evaluation work. This role centers on systems-level optimization, compiler/runtime topics, and creating clear rubrics for technical assessment.

This is a remote contract engagement for contributors located in the United States. The role requires reliable full-time availability (40 hours/week) and pays $70–$110 per hour.

• Employment type: Contractor (full-time, 40 hrs/week).
• Pay: $70 to $110 per hour (USD).
• Location: United States only; fluency in English required.

What you'll do

You will design performance-focused tasks, produce high-quality reference solutions, and evaluate submissions against clear technical criteria. Your work will make training data for optimization and runtime tasks both rigorous and teachable.

• Design challenging performance-engineering tasks spanning systems, compiler, and runtime topics.
• Write accurate, well-structured solutions demonstrating latency, throughput, or memory improvements.
• Evaluate task submissions and provide clear, constructive technical feedback.
• Develop and refine guidelines and detailed rubrics to assess systems design quality across AI workloads.
• Collaborate with subject-matter experts to keep training data consistent and accurate.

Requirements

Candidates must have hands-on experience demonstrating measurable production performance wins and be able to communicate technical results clearly in writing.

• Minimum 2+ years professional experience in performance engineering, systems programming, or low-level optimization.
• Deep, hands-on expertise in at least one of C++, Python, or Rust with production experience.
• Proven, demonstrable performance improvements (latency reduction, throughput gains, or memory footprint optimization).
• Experience writing technical evaluations, rubrics, or providing written feedback.
• Strong written communication and the ability to explain technical trade-offs clearly.
• Ability to work reliably 40 hours per week.

Helpful background

The following are not strictly required but will help you succeed and influence more advanced projects.

• Experience with compiler engineering, runtime performance, or low-level systems tuning.
• Familiarity with multiple systems languages and performance tooling.
• Prior work creating evaluation datasets or technical rubrics for engineering problems.

How this work is structured

You will create and review text-based tasks and evaluations (data type: text) and perform both generation and evaluation/rating work. Tasks focus on producing solutions and scoring technical quality rather than manual image/audio labeling.

As a contractor on OpenTrain, you will be assigned projects, deliver solutions and feedback according to rubrics, and iterate with reviewers to keep training data high quality.

• Label types: Text generation and evaluation/rating.
• Time commitment: Full-time, 40 hours per week (contractor).
• Work is remote; only US-based applicants will be considered for this listing.