A recruitment portal intended to eliminate racial discrimination and bias based on cultural backgrounds, and focus on skills to give equal employment opportunities.
Full demo with audio - https://www.youtube.com/watch?v=lS5pH5u_pTo
-
Technological viability: Our solution will grow with Watson. As Watson Speech to Text and Natural Language Processor improves over time, our solution will also improve.
-
Anticipated challenges: To add on to that, we are also limited to Watson's AI capabilities. To limit bias in the future, we would have to train Watson to recognize bias or create our own dataset and train a model to recognize bias. If we do not have that capability, we would want to have HR vet keywords that hiring managers can use, to minimize bias manually. Furthermore, Watson's error in recognizing certain words/phrases may be frustrating for some applicants, however this can be mediated by further training of text recognition.
-
Plan to overcome: While we have tackled a big part of bias in recruiting, we have yet to tackle bias in using Watson Natural Language processing to find keywords/phrases/concepts for hiring. This is something we intend to tackle along with minimizing the Watson error percentage in the interview process, whether that is by training Watson further or creating our own model to train, or even solving it manually, such as requiring HR to approve/vet keywords before going live.
DB_PW =
Using mongoDB - https://www.mongodb.com/cloud/atlas
API_TTS =
API_STT =
API_TA =
API_NLU =
API_PI =
API_TTS - https://www.ibm.com/cloud/watson-text-to-speech
API_STT - https://www.ibm.com/cloud/watson-speech-to-text
API_TA - https://www.ibm.com/watson/services/tone-analyzer/
API_NLU - https://www.ibm.com/cloud/watson-natural-language-understanding