OpenTrain is hiring a Machine Learning Engineering Evaluator for an AI training project. You will create, solve, review, and validate demanding machine-learning engineering tasks, then explain technical decisions, trade-offs, and limitations.
The work covers model development, training and inference systems, numerical computing, performance optimization, Python workflows, and technical evaluation.
- Develop and validate machine-learning models, training pipelines, inference systems, and supporting infrastructure.
- Implement model components, data pipelines, evaluation systems, and numerical methods.
- Build reproducible workflows with Python and command-line tools.
- Work with tensor operations, automatic differentiation, model architectures, tokenization, batching, and generation.
- Optimize latency, throughput, memory use, and hardware utilization.
- Diagnose numerical instability, tensor errors, memory bottlenecks, distributed-system failures, and performance regressions.
- Review AI-generated code and technical solutions.
- Design objective tests, benchmarks, and verification criteria.
What it pays and takes
Equivalent tools and substantial open-source or academic experience may also qualify.
- Pay: $100-$150 per hour.
- Time: Approximately 15 hours per week, with less than 20 hours per week expected.
- Schedule: Flexible, including the option to work weekends.
- Work setting: Fully remote contractor role.
- Location: remote, open to applicants in India and the United States.
- Language: Fluency in English.
- Education: Master's degree or PhD in computer science, machine learning, artificial intelligence, applied mathematics, statistics, engineering, or a closely related quantitative field.
- Experience: Strong professional or research experience in machine learning.
- Technical skills: Practical proficiency with Python and reproducible technical workflows.
- Frameworks and tools: Meaningful experience with at least two relevant machine-learning frameworks, libraries, or inference tools. Examples include PyTorch, JAX, NumPy, SciPy, SGLang, vLLM, llama.cpp, Hugging Face Transformers, and Hugging Face Tokenizers.
- Technical judgment: Ability to debug machine-learning systems beyond surface-level API use and clearly explain implementation choices, performance trade-offs, and failure modes.
How it works
You apply here in about two minutes; the application finishes on the hiring platform, with a short AI interview if you are shortlisted.
About AI training work
AI training and data-labeling work is the human side of building artificial intelligence. People prepare examples, review model outputs, test systems, and provide technical feedback so AI models can perform more reliably; experienced specialists are needed for complex engineering evaluation.
OpenTrain is a platform where people find AI training and data-labeling projects, build a profile, and apply for work. Creating an OpenTrain account is free.
