client logo
 

Organization

Curately AI

 

Legislation

NYC LL 144
 

AEDT Name

Search and Match

 

Distribution Date

20 Mar 2024

 

Audit Date

24 May 2024

 

Data Type

Historical

 

Data Period

1 Jan 2023 - 1 May 2024

 

Protected Attributes

EthnicityGender
 
Powered by
hai-logo
Summary of results
 

AEDT information

AEDT Description:

 
Curately AI leverages advanced machine learning techniques to transform both an individual’s resume/CV and job descriptions into detailed semantic representations. This process involves training neural networks on publicly available data and comprehensive job taxonomies to understand and accurately represent each skill. The platform generates a computed matching score by comparing the semantic representations of candidates' profiles with job requirements. This score reflects how closely an individual’s skills and attributes align with the needs of the specific job, without directly comparing candidates against one another. Additionally, recruiters can assign variable weights to different skills or attributes on a search-by-search basis, prioritizing those deemed most critical for the role.

Key Information:

 

While the primary function of Curately AI is to provide an accurate matching score using advanced artificial intelligence, it is crucial to recognize that this tool does not make or significantly assist in making employment decisions on its own. Curately AI specializes in analyzing and comparing candidates' profiles with job descriptions to generate a computed matching score that reflects the alignment of skills and attributes.

As part of our commitment to fairness and accuracy, Curately AI undergoes regular testing to identify and eliminate any potential biases in its matching algorithms. While the platform does not make employment decisions, ensuring that its scoring process is transparent and free from biases is essential for maintaining the integrity of the selection process.

Data Details:

 

Curately.ai provided historical data from applicants previously assessed by the system. The data file submitted contained 52,555 rows across 6811 jobs, with columns for company_name, job_position_name, user_id, gender, ethnicity, and user_job_score.

Original Sample Size: 52555
Number of Candidates with Unknown Characteristics: 756
Final Sample Size: 51799

Impact ratios

Group
Scoring Rate
Impact Ratio
Sample Size
American Indian Or Alaska Native 0.46 271
Asian 0.49 0.95 14359
Black Or African American 0.52 1 11781
Hispanic Or Latino 0.5 0.96 4896
Native Hawaiian Or Other Pacific Islander 0.49 135
Two Or More Races 0.46 0.89 3087
White 0.5 0.95 15676
— Group represents less than 2% of individuals, impact ratio is excluded from analysis.