Career Compass

• 

  Sample run for a Senior Data Analyst role. BT-coded mods seem to produce good scores.

• 

  As expected, QF and Math mods are less relevant to the role

• 

  Expanded requirement details

• 

  Full Breakdown of Job Requirements Met/Missed

• 

  Sample Mod Comparison

• 

  Option to Remove Requirements (if extras are picked up)

About the Project

As university students, we often find ourselves choosing modules without knowing how relevant they will be in the future. A module may sound interesting, but it can be difficult to tell whether it will help prepare us for the jobs we hope to pursue. At the same time, many graduates enter the workforce with valuable knowledge but still struggle to demonstrate how their education connects to what employers are looking for.

This uncertainty inspired me to build Career Compass.

Career Compass helps students connect their university modules, projects, internships, and resumes to real-world job descriptions. Students can enter the type of role they are interested in, load their NUS modules through NUSMods, and see which parts of their current learning already match the role. The app also highlights areas where they may benefit from additional projects or experience.

I wanted Career Compass to be more than just a score. It should help students make more confident decisions about what to learn next. For example, a student interested in becoming a data analyst can compare future modules and see which one best supports skills such as SQL, Python, data visualisation, cloud computing, or A/B testing.

Building the project taught me that creating a useful application involves much more than writing code. Real information is often messy. NUS module descriptions can contain long paragraphs, resumes can use very different layouts, and job descriptions include a mixture of responsibilities, qualifications, and general company information. I had to repeatedly improve how the application identified meaningful modules, requirements, projects, and internships.

One of the biggest challenges was making the results understandable. Similarity scores can feel abstract and discouraging if they are presented without context. I therefore redesigned the output into a clearer Career Compass Alignment Index, with covered requirements, critical gaps, and ranked future module recommendations. I also made sure that a flagged gap is presented as an opportunity for growth rather than a judgement of a student’s ability.

Through this project, I learned about natural language processing, product design, data quality, and the importance of testing ideas with the user experience in mind. Most importantly, I learned that technology can help students feel more in control of their academic and career decisions.

Career Compass is my attempt to make the path between university learning and employment easier to understand, one module choice at a time.

Built With

• hugging-face
• machine-learning
• natural-language-processing
• nusmods
• python
• scikit-learn
• sentence-transformers
• streamlit
• streamlit-community-cloud