Model-assisted evaluation and intelligent interview feedback

1. A method comprising:

storing, by a processing device, a trained modeling function, wherein the trained modeling function embodies a correlation between historical ratings and historical evaluation results of recorded responses that were responsive to a plurality of prompts presented to a set of persons, wherein the recorded responses are portions of digital interviews;

receiving, via a graphical user interface (GUI), a request to initiate a digital evaluation process for a second person being evaluated;

receiving recorded responses of the second person that are responsive to the plurality of prompts;

generating, by the processing device applying the trained modeling function to the recorded responses of the second person, a reviewing sequence by which to view the recorded responses of the second person, the reviewing sequence ordering the recorded responses of the second person from a most predictive prompt to a least predictive prompt of the plurality of prompts;

transmitting, by the processing device, data that causes display in the GUI of a list of the recorded responses of the second person in the reviewing sequence;

delivering, by the processing device to the GUI for review, one or more recorded responses listed first according to the reviewing sequence for the second person;

in response to receiving, via the GUI, one or more ratings for the one or more recorded responses, delivering, to the GUI, an evaluation decision prediction based on the one or more ratings; and

in response to receiving, via the GUI, an evaluation result associated with the evaluation decision prediction, ending, by the processing device, review of the recorded responses for the second person.

2. The method of claim 1 , further comprising generating, by the processing device applying the trained modeling function to a current rating of at least one of the recorded responses of the second person, the evaluation decision prediction for the second person.

3. The method of claim 1 , further comprising:

receiving an evaluation result and a current rating for at least one of the recorded responses of the second person;

inserting, into a training data set, the current rating and the evaluation result of the at least one of the recorded responses of the second person, to generate an updated training data set; and

updating training of the trained modeling function using a correlation between the historical ratings and the historical evaluation results of the recorded responses within the updated training data set, to generate an updated trained modeling function.

4. The method of claim 3 , further comprising:

generating, by the processing device applying the updated trained modeling function to the recorded responses of the second person, a second reviewing sequence by which to view the recorded responses of the second person; and

transmitting, by the processing device, second data that causes display of the list of the recorded responses of the second person in the second reviewing sequence.

5. The method of claim 1 , further comprising:

altering an order of remaining prompts in the reviewing sequence to provide an updated reviewing sequence; and

presenting, via the GUI, a first remaining response of the recorded responses of the second person according to the updated reviewing sequence.

6. The method of claim 1 , wherein the trained modeling function utilizes at least one of regression analysis, neural networks, support vector machines, decision trees, or a Markov model variant.

7. The method of claim 1 , further comprising determining whether an amount of the historical ratings and the historical evaluation results exceeds a threshold number of ratings and corresponding evaluation results before training the modeling function.

8. A computing system comprising:

a data storage device; and

a processing device, coupled to the data storage device, to execute instructions to:

store, in the data storage device, a trained modeling function, wherein the trained modeling function embodies a correlation between historical ratings and historical evaluation results of recorded responses that were responsive to a plurality of prompts presented to a set of persons, wherein the recorded responses are portions of digital interviews;

receive, via a graphical user interface (GUI), a request to initiate a digital evaluation process for a second person being evaluated;

receive recorded responses of the second person that are responsive to the plurality of prompts;

generate, by applying the trained modeling function to the recorded responses of the second person, a reviewing sequence by which to view the recorded responses of the second person, the reviewing sequence ordering the recorded responses of the second person from a most predictive prompt to a least predictive prompt of the plurality of prompts;

transmit data that causes display in the GUI of a list of the recorded responses of the second person in the reviewing sequence;

delivering, to the GUI for review, one or more recorded responses listed first according to the reviewing sequence for the second person;

in response to receiving, via the GUI, one or more ratings for the one or more recorded responses, delivering, to the GUI, an evaluation decision prediction based on the one or more ratings; and

in response to receiving, via the GUI, an evaluation result associated with the evaluation decision prediction, ending review of the recorded responses for the second person.

9. The computing system of claim 8 , wherein the recorded responses of the set of persons and the recorded responses of the second person comprise video recordings of responses to prompts in the digital interviews.

10. The computing system of claim 8 , wherein the processing device is further to receive a current rating of at least one of the recorded responses of the second person.

11. The computing system of claim 10 , wherein the processing device is further to generate, by applying the trained modeling function to the current rating of the at least one of the recorded responses of the second person, the evaluation decision prediction for the second person.

12. The computing system of claim 10 , wherein the processing device is further to:

receive an evaluation result for the at least one of the recorded responses of the second person;

insert, into a training data set, the current rating and the evaluation result of the at least one of the recorded responses of the second person, to generate an updated training data set; and

update training of the trained modeling function using a correlation between the historical ratings and the historical evaluation results of the recorded responses within the updated training data set, to generate an updated trained modeling function.

13. The computing system of claim 12 , wherein the processing device is further to:

generate, by applying the updated trained modeling function to the recorded responses of the second person, a second reviewing sequence by which to view the recorded responses of the second person; and

transmit second data that causes display of the list of the recorded responses of the second person in the second reviewing sequence.

14. The computing system of claim 8 , wherein the processing device is further to:

alter an order of remaining prompts in the reviewing sequence to provide an updated reviewing sequence; and

present, via a user interface, a first remaining response of the recorded responses of the second person according to the updated reviewing sequence.

15. The computing system of claim 8 , wherein the trained modeling function utilizes at least one of regression analysis, neural networks, support vector machines, decision trees, or a Markov model variant.

16. The computing system of claim 8 , wherein the processing device is further to determine whether an amount of the historical ratings and the historical evaluation results exceeds a threshold number of ratings and corresponding evaluation results before training the modeling function.

17. A non-transitory computer-readable storage medium including instructions that, when executed by a processing device, cause the processing device to perform operations comprising:

storing a trained modeling function, wherein the trained modeling function embodies a correlation between historical ratings and historical evaluation results of recorded responses that were responsive to a plurality of prompts presented to a set of persons, wherein the recorded responses are portions of digital interviews;

receiving, via a graphical user interface (GUI), a request to initiate a digital evaluation process for a second person being evaluated;

receiving recorded responses of the second person that are responsive to the plurality of prompts;

generating, by applying the trained modeling function to the recorded responses of the second person, a reviewing sequence by which to view the recorded responses of the second person, the reviewing sequence ordering the recorded responses of the second person from a most predictive prompt to a least predictive prompt of the plurality of prompts;

transmitting data that causes display in the GUI of a list of the recorded responses of the second person in the reviewing sequence; and

delivering, to the GUI for review, one or more recorded responses listed first according to the reviewing sequence for the second person;

in response to receiving, via the GUI, one or more ratings for the one or more recorded responses, delivering, to the GUI, an evaluation decision prediction based on the one or more ratings; and

in response to receiving, via the GUI, an evaluation result associated with the evaluation decision prediction, ending review of the recorded responses for the second person.

18. The non-transitory computer-readable storage medium of claim 17 , wherein the operations further comprise:

populating a ratings matrix with the historical ratings of the recorded responses of the set of persons, wherein each row of the ratings matrix comprises a vector of ratings for a person of the set of persons;

populating an evaluation results vector with the historical evaluation results corresponding to the historical ratings for the set of persons; and

training the trained modeling function comprising:

correlating a plurality of vectors of the ratings matrix with corresponding entries of the evaluation results vector for the set of persons; and

assessing prediction results for a plurality of possible sequences of columns of the ratings matrix.

19. The non-transitory computer-readable storage medium of claim 17 , wherein the operations further comprise:

receiving an evaluation result and a current rating for the at least one of the recorded responses of the second person;

inserting, into a training data set, the current rating and the evaluation result of the at least one of the recorded responses of the second person, to generate an updated training data set; and

updating training of the trained modeling function using a correlation between the historical ratings and the historical evaluation results of the recorded responses within the updated training data set, to generate an updated modeling function.

20. The non-transitory computer-readable storage medium of claim 19 , wherein the operations further comprise:

generating, by applying the updated modeling function to the recorded responses of the second person, a second reviewing sequence by which to view the recorded responses of the second person; and

transmitting second data that causes display of the list of the recorded responses of the second person in the second reviewing sequence.