Researcher Mathematics (Coding)

Company description

Dusker AI specializes in evaluating and benchmarking advanced AI systems through rigorous, research-driven methodologies. We partner with organizations developing large language models, AI agents, and autonomous systems to measure performance across reasoning, reliability, adaptability, safety, and domain expertise.

Our evaluation frameworks are designed by subject matter experts and researchers, enabling deeper assessment than traditional benchmarks. By identifying strengths, uncovering failure modes, and testing real-world capabilities, Dusker AI helps organizations build AI systems that are robust, trustworthy, and deployment-ready.

We work across a wide range of domains including mathematics, physics, chemistry, biology, software engineering, and applied sciences, creating high-quality evaluation datasets and benchmark suites that drive the next generation of AI performance.

Role description

Researcher – Mathematics (Coding)

Compensation: $20–$25 per accepted task

Location: Remote

Education Requirement: Master's or PhD

Publication Requirement: Peer-reviewed publications, conference papers, or equivalent research contributions preferred

Application Requirement: Applicants are strongly encouraged to include links to their Google Scholar profile, ResearchGate profile, ORCID profile, arXiv publications, personal academic website, or other verifiable research publications within their CV/Resume.

Applications with demonstrated research experience and accessible publication records will be prioritized during the review process.

We are seeking a highly analytical Mathematics Researcher with strong programming expertise to contribute to the design and evaluation of advanced AI systems. This remote position focuses on developing mathematically rigorous benchmarks, evaluating model reasoning capabilities, and conducting research at the intersection of mathematics, computation, and AI.

The successful candidate will formulate challenging mathematical tasks, implement algorithms and simulations, analyze model behavior, and develop evaluation methodologies that measure the accuracy, robustness, and reasoning abilities of AI systems.

Key Responsibilities

• Design mathematically rigorous benchmark tasks and evaluation frameworks for AI systems

• Develop coding-based mathematical problems, simulations, and computational experiments

• Analyze model outputs and identify strengths, weaknesses, and failure patterns

• Conduct quantitative research on reasoning, problem-solving, and algorithmic performance

• Collaborate with multidisciplinary teams to improve benchmark quality and coverage

• Document methodologies, research findings, and evaluation results

• Contribute to technical reports, publications, and internal research initiatives

• Stay informed about developments in AI evaluation, AI safety, mathematical reasoning, and machine learning

Qualifications

Required:

• Master's or PhD in Mathematics, Applied Mathematics, Computer Science, Physics, or a closely related quantitative discipline

• Demonstrated research experience with peer-reviewed publications, conference papers, or equivalent scholarly contributions

• Strong foundation in pure or applied mathematics, including formal reasoning, proof construction, and quantitative analysis

• Proficiency in Python and experience developing algorithms, simulations, or analytical tools

• Experience with statistical analysis, experimental design, and data interpretation

• Excellent written and verbal communication skills in English

• Ability to work independently in a remote research environment

Preferred:

• Experience evaluating AI systems, machine learning models, or large language models

• Background in mathematical optimization, probability, statistics, numerical methods, or computational mathematics

• Experience developing benchmarks, datasets, or evaluation frameworks

• Familiarity with AI safety, model alignment, or AI benchmarking methodologies

Candidates with strong academic research backgrounds, publication records, and demonstrated expertise in mathematical problem solving are especially encouraged to apply.