Soft skill assessments date back 50+ years - we've created a gamified assessment to determine someone’s potential fit or success in a role and mitigate bias.

pymetrics is converging modern cognitive science with cutting edge AI to accurately and scalably assess people to jobs

It Starts With Soft Skills

Cognitive science promotes equity + flexibility in the future of work

Advancements in cognitive science led to two primary differentiators for pymetrics soft skills:

Equity: Unlike traditional tools, pymetrics' soft skills are evenly distributed across gender, race and socioeconomic status, making them the key to equitable soft skill assessment.
Flexibility: pymetrics' soft skills platform is fit-based, understanding inherent fit to role rather than using a one-size-fits all approach or restrictive credentials.

Audited, fairness-optimized AI

We can design AI to be both accurate and objective by using these three principles. First, we start with better data: evenly-distributed soft-skills data. Second, we only release algorithms if they meet stringent accuracy and unbiased standards (lacking disparate impact). Finally, we don’t ask you to take our word for it: we had our platform independently audited to ensure full transparency.

See full report here.

Read about our audit

The standards we follow

UGESP

Each step of the pymetrics process has been implemented in compliance with the Uniform Guidelines on Employee Selection Procedures (UGESP). This includes job analysis, criterion-related validation (concurrent validity and predictive validity), reliability, and fairness.

EEOC

While the Equal Employment Opportunity Commission (EEOC) does not rubber-stamp assessment products as compliant or non-compliant with its guidelines, each step of the pymetrics process was designed with UGESP in mind and is implemented using methods that adhere to these regulations.

Download compliance whitepaper

A broader community

pymetrics is committed to contributing to scientific knowledge about ethical AI. We work with academics from a variety of disciplines including computer science, psychology, economics, and law, to provide new information about our technology that is transparent and useful for diverse audiences.

Current academic collaborations

Christo Wilson

Computer Science Professor at Northeastern

A third-party audit of the pymetrics model-building process, including a controlled experimental of the de-biasing capability.

Arun Sundararajan

Technology Professor at NYU

An analysis of how pymetrics affects the socioeconomic diversity of candidates who are passed through the first stage of hiring pipelines.

Ifeoma Ajunwa

Law Professor at Cornell

A comprehensive ethnography of pymetrics as an AI-based hiring assessment, in support of a National Science Foundation grant.

David Autor

Economics Professor at MIT

A comparative analysis of the prospects for mitigating bias in human decision-makers versus algorithms.

Lori Foster

IO Psychology Professor at NC State University

An exploration of how pymetrics can be used to approximate various psychometric estimators, without the bias of traditional assessments.

Science.

It Starts With Soft Skills

Audited, fairness-optimized AI

Backed by research

MIT - Cognitive Science as a New People Science for the Future of Work

ACM FAccT 2021 - Building and Auditing Fair Algorithms: A Case Study in Candidate Screening

UMass - Using Technology to Increase Fairness in Hiring

In the news

Brookings

Auditing employment algorithms for discrimination

Fast Company

Seven simple steps to design more ethical AI

HBR

Using AI to Eliminate Bias from Hiring