beyond-resume

A Machine Learning Based Approach to Hiring Qualified Individuals
This implementation is something which profits both the employer and the applicant. Below are some the key highlights of the implmentation -

• MongoDB based backend provides seamless and fast querying and information retrieval
• An enhanced job-descrition to cater other possible areas of expertise which might be closely linked to the ones mentioned in the description.
• A comparative measure of applicants abilities and skills with the requirements of the position.
• A Flask based web application which has been integrated with both Python and MongoDB.

Data Gathering, Pre-processing and Backend Development

The data was gathered for 4 types of job description -

• Data Scientist
• Data Engineer
• Software Engineering
• Quantitative Analyst

Some key steps involved in this process include -

• The job descriptions were filtered for keywords and their related vocabulary through web-scrapping
• The data was stored in the form of a Graph where each node is linked with its related vocabulary words e.g. Deep Learning is linked with Neural Networks

Applicant Evaluation and Front-end Development

Using the data gathered, we perform the following measures of evaluation of an applicant -

• Based on the applicant skills and how closely it is related to the skills identified from the job description, we provide a score to their skills through topological measures of network analysis
• The top 15 skills of the applicant which align with the job description are returned Skills pertaining to the Data Science Job Description
  

Applicant Friendly Features

• The web application lets the applicant know if there is a very low match between his expertise and the job's required skill-set.
• Based on the applicant's skills, the application is capable of suggesting other jobs which are in the system which are a better match to their skill-set

Conclusion

Some of the key points which can be inferred from our development are -

• The development is highly scalable with its ability to improve based on the dataset used to create the backend using NLP and Textual data mining techniques.
• The backend for this application can be made lighter with the envolvement of AI based recommendation system

Built With

• azure
• beautiful-soup
• flask
• fuzzymatching
• graph-mining
• html
• mongodb
• networkx
• python