Machine Learning can take as input resumes and output either to email the candidate or not
My question is : Previously on the first week we saw that machine learning works well when learning simple concepts , anything that can be done with one second of thought and I believe that observing a resume might take some time, so how is that possible?
Current state of the art is able to process complex inputs like natural language without problems. The real thing here is if this kind of systems actually contribute to making recruiting process better.
Anyway, NLP has improved a lot in this recent years. Check GPT-3 and LaMDA models from OpenAI and Google respectively.
Resume crawling is just the tip of the iceberg of what companies can already do with user submitted text and machine learning. Not an endorsement, and I believe the product offering might have been deprecated, but check out the feature summary of IBM Personality Insights, for one example…
Detailed personality portraits - Use linguistic analytics to infer individual personality characteristics (like big five, needs, and values) from digital communications such as email, blog posts, tweets, and forum posts.
It produced output like this…
Hello @Bilel_Djemel , Welcome to DeepLearning.AI
Thanks for your post.
As per your query, it is possible to train and build efficient systems for resume selection. As far as I know, there are companies which are presently using them. So according to the needs of your organisation and skills required for the job opening, you can make and train a model with some of the possible input features being skills, projects ( tech stack used in them ), coding background ( if platform link given , evaluate ratings ), Prior work experience. In addition to this, you can also provide or sent out forms to gain extra information to train your model.
But there are some challenges faced like biasness that can arise in this process. Biased Model can lead to biased resume selection as also mentioned in AI For EveryOne course. But that too can be handled.
I hope that I have answered your question. Happy Learning !
Amit Shukla