| Group |
2020-04 |
Status |
completed |
| Title |
HR platform – Path to diversified workforce |
| Supervisor |
Yan Liu |
| Description |
Todays recruitment market is creating multiple challenges for both employers and potential employees. The ability to find the right candidates is critical for any organization, the cost of hiring is high, if wrong candidates are selected, and it creates significant challenges for the organization. On average, the hiring of the candidate and training will carry the cost of a yearly salary.
On the other hand, future employees are often concerned that due to bias and prejudice, they may not be considered, and their name, race or other factors will serve to their disadvantage. Process is often complicated, unfair and bias. We would like to address this friction point.
We would like to create a platform that identifies/controls for bias, prejudice and, at the same time, recognizes the best candidate for a job. The goal would be to utilize the technology to address the two main challenges that employees and employers face today.
CGI strongly believes that this platform shall be available to everybody, and it will ask to build utilizing open-source technologies and design as an open source solution available to any organizations that would like to use it and implement it.
CGI will provide weekly supervision with senior resources and under supervision of Director/VP resources allocated to work with students. CGI will provide help in both design and tech if required. We are looking to work with the brightest students at Concordia for this initiative. We believe that this project will require a very diversified team that can help build this solution.
Platform
We would like students participating in the capstone project to build the recruitment platform that would remove any bias and prejudice from the recruitment process and aggregate the information available today in multiple sources.
Key functionalities:
Web & Mobile Application Front-End– we are looking for UX utilizing best in class human-centric and material design principles. ML & database should communicate with front-end through API to extract data from resume.
Database – all the information shall be stored, and safety and security shall be considered as a priority. The information would include the resume files and cover letters, and the information will be categorized to enable the analytics capabilities for the super-platform users. GDPR shall be considered, and the user shall have access to remove its information. Extracted data should live in NoSQL or SQL database based on Capstone team recommendation.
Algorithm – Match me algorithm will need to be developed to identify the right candidacy with particular posting. This tagging system would entail Natural Language processing, perhaps in a Transformer architecture. Tags will need to base stored in above database structured for use in API.
Recommendation engine – Based on resume the engine shall be able to recommend best job matches and based on job description should recommend top 5-10 candidates. Nearest Neighbor and Collaborative Filtering are two potential approaches.
Analytics – Dashboard and other insight and trends on aggregate analytics will need to be enabled, via api output that can allow connection to external analytics platform such as PowerBi via JSON/XML output.
Connectivity – Platform shall have API connectivity to other recruitment platforms, i.e.indeed, API should be written on Open API industry standard.
|
| Requirement |
Some background in programming – potential skills include Python, JavaScript frameworks such as React, Vue, Next, Flutter, API design, Tensorflow. Any skills not currently present will be coached by CGI team. |
| Tools |
We would like our capstone team to have the experience of working in a real-world enterprise environment. Thus we will expect standard software delivery tools to be used. These include:
Version Control in Git (GitHub or GitLab)
Ci/Cd in Gitlab or Jenkins
Agile Project Management (in Trello or other)
Dev/Production Containers in Docker
Container Orchestration in Kubernetes
MLOps in MLFlow or KubeFlow
All of the above areas best practices will be coached by CGI team. |
| Number of Students |
5 |
| Students |
Dragos Alex Badiu
Ziad Bahloul
Manar Alchirazi Alsabbagh
Felix Wawrosz:
|
| Comments: |
Full - stack, front-end, UI UX designers, Machine Learning programmers/enthusiasts |
| Links: |
|