AWS Trainium/NKI (TRN) AI Task Auditor - Freelance AI Trainer Project
About this role
Evaluating AWS Trainium and Neuron Kernel Interface tasks forms the daily focus of this freelance project, ensuring that training workflows for artificial intelligence systems remain technically rigorous and practically sound. Specialists audit tasks for accuracy, realism, and reproducibility while identifying hardware acceleration inefficiencies and execution bugs.
Meridial runs this global remote initiative to source technical experts who can troubleshoot complex scenarios and deliver actionable feedback. Candidates who possess deep knowledge of AWS Trainium architecture and strong analytical abilities fit well into this auditing project.
What you will do
- Assess whether AWS Trainium and NKI tasks are technically accurate, realistic, solvable, and reproducible.
- Verify that evaluation tasks are supported by reliable tests and appropriate criteria.
- Provide clear and actionable feedback on identified hardware acceleration inefficiencies.
- Detect kernel execution bugs and logic errors within the evaluated tasks.
What you need
- Demonstrate professional experience and deep technical knowledge in AWS Trainium architecture.
- Bring practical expertise in model compilation and accelerator optimization.
- Possess familiarity with Neuron Kernel Interface frameworks.
- Apply strong analytical and problem-solving skills to troubleshoot complex technical scenarios.
Skills and tech stack
Questions candidates ask
What is the remote work policy for this role?
This position is fully remote and available worldwide. As a contractor, you will need to supply your own secure computer and high-speed internet.
What is the pay rate for the project?
The project offers a pay rate of $60 per hour. The exact rate is determined after evaluating your experience, expertise, and geographic location.
What kind of tech stack and tools are used?
The project focuses on AWS Trainium and Neuron Kernel Interface frameworks. Work involves model compilation and accelerator optimization.