Singapore’s healthcare system, the backbone of its economic success, is tethering on the edge of a crisis - a critical shortage of nurses. 

This comes in a time of great uncertainty when Singapore is being plagued by both a rapidly ageing population and a declining birth rate.

In 2022, there were a total of around 43000 nurses across the island. However, considering Singapore’s rapidly ageing population, MOH projects this number to rise by 24000 by 2030.  

In the face of this crisis, we present an innovative and game-changing platform that seamlessly connects passionate nurses with healthcare facilities in need. 

Brace yourself as we navigate through a digital solution that not only addresses the shortage of nurses but also transforms key challenges into opportunities.

Let’s start by looking at the workflow, to get an overview of the different parts and functions of our website. Firstly, nurses can sign up and key in their details, while Farrer Park Hospital can post the jobs as well as the job requirements. The price of the job will be decided by our price optimization feature.

The nurses would then be able to search and apply for these jobs. Our AI-assisted nurse selection would rank the nurses, before a nurse is selected for his/her job. Lastly, the final rating of the nurse will be given by the nurse’s supervisors, and this additional data would be stored and used to rank them for other jobs.

With the help of this freelance website that matches Farrer Park hospital’s needs to the nurses, this nurse shortage in Singapore will be heavily mitigated.

Let’s talk about our architecture behind our web app. We used static site generation(SSG) to render our dynamic website frontend into static files stored in the scalable S3 Storage. These static pages communicates with backend using REST APIs to our flask application which is hosted on Elastic Beanstalk. The flask application can then retrieve any nurse or job data from the DynamoDB we used to store our data and if needed can send the data to our AI model which is deployed on sagemaker to predict the score. All these data is then sent back to frontend where it is manipulated and displayed

So let’s focus on the AI-Assisted Nurse Selection process When a nurse applies for a job, it will send the nurse particulars to the backend’s machine learning model. Our model is trained using ensemble methods by combining multiple weak models to get a strong robust model. This model will then predict a score for that nurse on how well they would probably do for that job and then sends the predicted score back to front end. We experimented with diff deep learning and ensemble learning models and found that the xgboost(extreme gradient boosting) performs best for our task.

The frontend will then manipulate the data and present it where the top 30% of nurses will be displayed on admin page and recommended to admin for review and hire.We also have a price optimisation feature where the salary is dynamically updated in real time based on certain factors such as in this case, the popularity of that job and the the urgency of it. for example the closer the deadline of the job the higher the salary will be in order to attract the local talent and fill in the vacancies for the job. Next, let me take u through what our actual site looks like for both the hospitals and the nurses

Dashboard to view jobs currently listed Able to check out the applicants for the specific jobSimple job creation page, Just have to fill in the general details (point to each datapoint)For the nurses,Their own dashboard to view their earnings month on month, as well as past jobs taken for each time periodUnder jobs tab, can view the jobs that are available to apply for Drop down for details Upon clicking on the view details icon, more detailed page shows up with more information on the job Moving on, lets talk about our future works.

Firstly, we plan to develop a progressive web-app using our existing web-based code. We recognise the convenience and efficiency of using a mobile app. Essential features like push notifications, can be used to notify nurses and admin staff of new job applications and listings.

We have even envisioned the user interface of our app using Figma, and here are our designs for the hospital side, and the nurse side.

For the hospital side , we have curated the Sign In, Register as well as the Terms and Conditions page. Moving past that, this is the nurse overview, and the page to create a new job listing. For nurses , we have also reimagined the nurse home page to make it more mobile-friendly.Weighted Bipartite MatchingAble to find direct match, no need for top 30% most suitable candidates reduces admin work, as hospital will get the optimum person

Replace existing program with a simple AI (like logistic regression)More data collected from the app to train AIAble to make better decision