Efficient percentile estimation for applicant rankings

A system comprising:

• one or more computer processors;
• and
• one or more computer readable mediums storing instructions that, when executed by the one or more computer processors, cause the system to perform operations comprising:
• gathering historical feedback data describing previously executed search queries, the previously executed search queries being executed in relation to previously available opportunities, the historical feedback data including profile feature data describing profiles returned as search results to the previously executed search queries, opportunity feature data describing the previously available opportunities, as well as interaction feature data describing interaction with the profiles returned as search results to the previously executed search queries;
• generating a machine learning model based on the historical feedback data, the machine learning model configured to output a probability score indicating a strength of a candidate for a target opportunity based on input of profile feature data describing a respective profile of the input candidate and opportunity feature data describing the target opportunity;
• obtaining parameters associated with a query for a ranking of candidates for a first specified opportunity, wherein the parameters comprise data identifying at least a first candidate and the specified opportunity;
• generating an input based on profile feature data describing a profile of the first candidate and opportunity feature data describing the specified opportunity;
• providing the input into the machine learning model, yielding a probability score indicating a strength of the first candidate for the specified opportunity;
• obtaining a fixed number of quantile thresholds associated with the parameters, the fixed number of quantile thresholds previously calculated from a distribution of probability scores for the candidates;
• using the obtained quantile thresholds and the first probability score, performing an interpolation technique to derive an estimated quantile for the first probability score;
• and
• presenting, via a user interface, the estimated quantile for the first probability score indicating a position of the first candidate within the ranking.

A non-transitory computer readable medium storing instructions that, when executed by one or more computer processors of one or more computing devices, cause the one or more computing devices to perform operations comprising:

• gathering historical feedback data describing previously executed search queries, the previously executed search queries being executed in relation to previously available opportunities, the historical feedback data including profile feature data describing profiles returned as search results to the previously executed search queries, opportunity feature data describing the previously available opportunities, as well as interaction feature data describing interaction with the profiles returned as search results to the previously executed search queries;
• generating a machine learning model based on the historical feedback data, the machine learning model configured to output a probability score indicating a strength of a candidate for a target opportunity based on input of profile feature data describing a respective profile of the input candidate and opportunity feature data describing the target opportunity;
• obtaining parameters associated with a query for a ranking of candidates for a specified opportunity, wherein the parameters comprise data identifying at least a first candidate and the specified opportunity;
• generating an input based on profile feature data describing a profile of the first candidate and opportunity feature data describing the specified opportunity;
• providing the input into the machine learning model, yielding a probability score indicating a strength of the first candidate for the specified opportunity;
• obtaining a fixed number of quantile thresholds associated with the parameters, the fixed number of quantile thresholds previously calculated from a distribution of probability scores for the candidates;
• using the obtained quantile thresholds and the first probability score, performing an interpolation technique to derive an estimated quantile for the first probability score;
• and
• presenting, via a user interface, the estimated quantile for the first probability score indicating a position of the first candidate within the ranking.