Machine Learning Research Proposal Specialist

Overview

We are looking for a Machine Learning Research Proposal Specialist to create a polished, professional research-style write-up for an ML evaluation task. The submission should propose a machine learning problem within the writer’s domain of expertise and must be directly related to machine learning. The final deliverable should be written in markdown format, follow the provided template exactly, and be no longer than 1–2 pages with a maximum of 8 paragraphs.

Project Goal

The goal of this project is to design a machine learning benchmark or problem that is complex enough to challenge a strong state-of-the-art LLM. The problem should not be a simple generic ML pipeline task. It should require deep domain expertise, advanced reasoning, and layered technical understanding that would be difficult for a general-purpose model to solve without specialized knowledge.

Responsibilities

The selected specialist will write a complete research proposal-style submission using the required sections: Domain and Motivation, Task Design, Data and Data Generation, Sources of Difficulty, Reference Solution and Scoring, and Citations. Each section should be concise, polished, and limited to one paragraph. The proposal should clearly explain the ML domain, why the task is meaningful, what is being predicted, how the data would be generated or adapted, why the task requires extrapolation, and why naive approaches would fail.

Ideal Candidate

The ideal candidate has strong experience in machine learning, applied AI, data science, or ML research, with expertise in at least one technical or scientific domain such as healthcare, finance, robotics, cybersecurity, biology, materials science, geospatial analysis, NLP, computer vision, or another complex field. They should be able to design a creative ML task that rewards real domain knowledge rather than basic textbook ML skills.

Required Skills

Strong understanding of machine learning problem formulation, benchmark design, data generation strategies, train/test split methodology, model evaluation, and scoring metrics is required. The candidate should also be comfortable explaining why a task is difficult, what domain-specific insights are needed, and how an expert would approach solving the problem. Clear technical writing and the ability to produce concise markdown-formatted documentation are essential.

Deliverable

The final deliverable will be a markdown-formatted research proposal following the provided template. It should be professional, concise, and no more than 8 paragraphs total. The proposal must involve machine learning directly and should include citations where applicable, such as relevant research papers, datasets, GitHub repositories, or other credible references that support the proposed task.

Important Notes

This is a machine learning evaluation project, so proposals that are not directly related to ML will not be accepted. The model output should be ignored during the evaluation process, and once the prompt has been entered and the model generates a response, the evaluator should select “Done Chatting.”