Background

Prentice is a company review and salary information platform for internship seekers. There are many job review platforms out there, but none cater towards the needs of would-be-interns. As a temporary employee, it’s crucial for interns to pick the right company and project to inform their career decision making. Prentice’s goal is to empower better decision making so students can figure out their ideal career and work environment more quickly.

The current method students use to gain information regarding a company’s internship experience is through upperclassmen and peers who have gone through the same experience. Unfortunately, it’s hard to figure out if you’re being compensated fairly and whether a company’s projects are suitable for interns to learn, when you don’t have connections to the companies you want to work at. Not to mention the limited data points you get by asking people on your personal network: not enough to make good, informed decisions.

As current students ourselves, we know how hard it is to find relevant information for that dream internship. Through Prentice, we want to provide a solution for our own problems and for many other students, so they can pick the best place to start their career.

Execution

In a short span of time, we managed to write relatively clean code and complete numerous features. Regarding our work for each learning path:

• Machine Learning: Developing sentiment analysis model via transfer learning and served to API, implementing keyword extraction algorithm, and building review recommendation system model using cosine similarity that can be integrated to backend.
• Mobile Development: Implement account authentication using firebase, and also accommodate users to be able to use the application by creating a UI interface, which is connected to the API using the retrofit library. We also integrated a machine learning system to the frontend that has been integrated with the cloud.
• Cloud Computing: Implemented Infrastructure-as-Code (IaC) using Terraform, serverless computing, and various other cloud technologies. We used Cloud Run to deploy our backend app. In addition, we used a custom logger instance to log events in the app and include other details such as timestamp, caller method, etc. We utilized many cloud services such as GCS, Cloud Run, Artifact Registry, Cloud SQL.

Built With

• cloud-run
• devops
• fastapi
• kotlin
• postgresql
• python
• recsys
• sentiment-analysis
• sklearn
• terraform