ML Research Engineer, Inference and GPU Kernels

You will create realistic technical scenarios and problem sets about machine learning inference and GPU kernel engineering. You will also write reference solutions and grading rubrics that set clear standards for correctness, depth, and technical judgment.

• Review AI-generated answers for accuracy, reasoning quality, and understanding of the technical field.
• Give structured feedback and write explanations that make expert judgments clear.
• The focus is evaluating and explaining technical work, rather than production engineering.

Requirements and availability

• The role requires 6 to 10 years of relevant professional experience in machine learning inference, GPU kernel engineering, inference serving, kernel optimization, or large-model deployment infrastructure.
• Bring current or recent hands-on experience in the field, a doctorate, or a master's degree backed by a strong research record.
• Strong written communication and the ability to work independently, asynchronously, and remotely are required.
• Prior research publications, open-source contributions, or recognized work products are useful. Experience evaluating, reviewing, grading, editing, or conducting code or peer review can strengthen your fit.
• Familiarity with SGLang, vLLM, Mamba or Mamba2, TensorRT-LLM, or GPU kernel work is helpful.
• The work is remote, flexible, asynchronous contract work, with a commitment of 20 or more hours per week.
• Location: Open to candidates in Albania, Algeria, Antigua & Barbuda, Armenia, Australia, Austria, the Bahamas, Bahrain, Belgium, Benin, Bhutan, Bosnia & Herzegovina, Botswana, Brunei, Bulgaria, Canada, Costa Rica, Côte d’Ivoire, Croatia, Cyprus, Czechia, Denmark, Ecuador, El Salvador, Estonia, Ethiopia, Finland, France, the Gambia, Germany, Greece, Guatemala, Guyana, Hong Kong SAR China, Hungary, Iceland, India, Indonesia, Ireland, Israel, Italy, Jamaica, Jordan, Kenya, Kuwait, Latvia, Liechtenstein, Lithuania, Luxembourg, Madagascar, Malaysia, Malta, Mauritius, Mexico, Moldova, Mongolia, Morocco, Mozambique, Namibia, the Netherlands, New Zealand, North Macedonia, Norway, Oman, Pakistan, Panama, Peru, the Philippines, Poland, Portugal, Qatar, Romania, Rwanda, Senegal, Serbia, Singapore, Slovakia, Slovenia, South Africa, Spain, Sri Lanka, St. Lucia, Sweden, Switzerland, Taiwan, Tanzania, Thailand, Trinidad & Tobago, Tunisia, Türkiye, the United Arab Emirates, the United Kingdom, the United States, Uzbekistan, and Vietnam.
• Other eligible country codes: MY, MZ, NA, NL, NO, NZ, OM, PA, PE, PH, PK, PL, PT, QA, RO, RS, RW, SE, SG, SI, SK, SN, SV, TH, TN, TR, TT, TW, TZ, US, UZ, VN, ZA.

Pay

• Pay is $100 to $170 per hour.

How to apply

Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI training work

AI training work is the human side of building AI: people prepare examples and review model responses so systems can learn from them. Experienced specialists are paid to bring their subject knowledge and judgment to that work.