machine learning engineer for LLM reasoning

Alignerr works with leading AI research labs on technically challenging problems in artificial intelligence, including the training and evaluation of cutting-edge LLMs.

• Author complex, high-fidelity reasoning traces that document how an LLM should plan, use tools, and reach decisions for sophisticated technical tasks;
• Design structured traces capturing multi-step problem decomposition, intermediate reasoning, and decision logic;
• Review and refine traces for consistency, logical soundness, and alignment with LLM training objectives;
• Develop data strategies that help models navigate ambiguous, multi-constraint, real-world scenarios;
• Contribute architectural insight to trace design, including why and how a model should reason through tasks.

• Strong understanding of model reasoning, evaluation, and training dynamics;
• Experience as an ML practitioner;
• Ability to decompose complex technical problems into clear, logical, well-documented steps;
• Familiarity with LLM behavior, model failures, and high-quality training data;
• Detail-oriented and rigorous approach to clarity and correctness;
• Ability to work independently and produce high-quality results;
• Nice to have: prior experience with data annotation, data quality assurance, or model evaluation pipelines, background in RLHF, chain-of-thought prompting, or reasoning-focused ML research, Grandmaster or Master-level Kaggle competition results, experience building or evaluating agentic AI systems with tool use and multi-step planning.

• Hourly contract with a commitment of 10–40 hours/week;
• Fully remote and async-friendly work;
• Freelance autonomy;
• Direct work with leading AI research labs;
• Exposure to cutting-edge LLM training and evaluation;
• Potential for ongoing contract extension as new projects launch.