Applied Mathematics Intern

Voushify is building a professional trust platform centered on verified recommendations and professional interactions. A core part of the platform is the Trust Score — a quantitative framework designed to translate different trust signals into a meaningful, transparent, and defensible measure of professional credibility.
We are looking for a mathematically strong intern to help us refine, test, and validate the Trust Score methodology.
What You’ll Work On
You will work directly on the mathematical framework behind Voushify’s Trust Score, including:

• Review the existing Trust Score methodology and identify mathematical weaknesses, biases, and edge cases.
• Develop and test alternative weighting and scoring models.
• Determine how different trust signals should contribute to an overall score.
• Model factors such as the quantity, quality, recency, diversity, and strength of professional recommendations.
• Explore methods for preventing score manipulation, including reciprocal recommendations, coordinated behavior, and artificial score inflation.
• Determine how confidence should change when a user has limited versus substantial evidence.
• Explore normalization techniques so scores remain comparable across users with different professional histories.
• Run simulations and sensitivity analyses to understand how the score behaves under different scenarios.
• Help establish mathematical rules for how the Trust Score evolves as new information is added.
• Document the methodology so the scoring system is explainable to both technical and non-technical stakeholders.
  Key Questions You May Help Solve
  For example:
  Should 20 recommendations from similar people be worth more or less than 8 recommendations from highly diverse professional relationships?
  How much should an older recommendation decay over time?
  How should we distinguish between a score of 85 based on five observations and an 85 based on fifty?
  How do we prevent users from gaming the system while keeping the scoring methodology understandable?
  What mathematical evidence is required before we can confidently say one professional profile has stronger trust signals than another?
  Ideal Candidate
  We are particularly interested in students studying:
• Mathematics
• Applied Mathematics
• Statistics
• Operations Research
• Data Science
• Quantitative Economics
• Computer Science with a strong mathematical background
  Strong candidates should be comfortable with probability, statistics, mathematical modeling, optimization, and quantitative reasoning.
  Python experience is strongly preferred, particularly NumPy, pandas, SciPy, or similar quantitative tools.
  Knowledge of Bayesian statistics, graph/network analysis, reputation systems, fraud detection, or machine learning is a plus but not required.
  What Success Looks Like
  The goal of the internship is not simply to produce another algorithm.
  By the end of the project, we want to have a mathematically defensible Trust Score framework with clearly defined inputs, weights, confidence levels, safeguards against manipulation, and documented reasoning behind the model.
  You will have the opportunity to help design a quantitative system that could become a foundational part of how Voushify measures professional trust.
  Why This Internship Is Different
  This is an applied mathematics role rather than a traditional software-development internship. You will be working on an open-ended problem where there may not be one mathematically “correct” answer.
  Your work can directly influence the architecture of a real product and the way professional trust is represented on the platform.