IBM Watson Recruitment
Team Size: 7
Team Member(s): Joydeep Mondal, Sudhanshu Shekhar Singh, Ritwik Chaudhuri, Manish Kataria, Kushal Mukherjee, Gyana Parija
Time Period: July'16-Present
Technologies/Concepts Used: PySpark, Kafka, Cloudant, ObjectStore, Theano, LSTM, SOAR, AnyLogic
IBM Watson™ Recruitment is a cognitive talent management solution that increases recruiter efficiency to allow HR to improve and accelerate people’s impact on the business.
It automatically predicts, without bias, the best suited candidates who are most likely to succeed in an organization.
In my current job role I work on both research and development aspects of it’s cognitive component -
- Machine Learning Pipeline Design - Designing the learning and real-time multi-client classification architecture pipeline for predictive analytics on kafka and pyspark.
- Dynamic Taxonomy Generation - Implementing a LSTM based architecture for generating taxonomies from a dataset of skill names. (Upcoming Paper Submission)
- Job Similarity Computation - Exploiting the semantic and syntactic structure of job description documents to find job similarity.
- Collaborative Cognition - Developing a knowledge exchange framework for different stakeholders in the recruitment process. Modelled as independent cognitive agents, each stakeholder gets to influence and get influenced by other agents, eventually leading to a more adoptible list of candidates. (Upcoming Paper Submission)