Borderless

Dreams, ambitions, and education all share one thing in common: no borders. At Borderless, we aim to help international students immigrate to achieve their borderless dreams.

Comment

Inspiration

We believe that education, just like dreams, should exist without borders. For international students, the opposite is the reality.

Faced with financial, legal, and personal borders, the pathway to education and success can often seem daunting. Adding on to this problem is immigration agencies that take advantage of international students - often charging them higher prices and sometimes resulting in fraudulent applications.

As such, we wanted to make an application that would make the immigration process easier for international students and hence came Borderless!

What it does

Borderless is an immigration app designed specifically for international students. It provides a platform where users can access comprehensive information about immigration procedures, visa requirements, job opportunities, and cultural integration tips through checklists.

These checklists offer personalized action items based on the user's profile, such as what documents they need to gather, what fees they need to make, suggested courses, and international language proficiency goals.

Our current application only supports checklists for one type of visa: F-1 Visa.

How we built it

Borderless is based on NextJS and Python. We set up a FastAPI server that accesses the Gemini AI from the Google AI API service. Next, we set up a frontend website with the NextJS and React (Inspired from: https://github.com/themefisher/bigspring-light-nextjs).

For our logic, we implemented a cheap version of RAG (Retrieval Augmented Generation). Instead of collecting various documents and making a vector database, we had a local version of the documents and then fed into the LLM. In addition, we used Agentic AI approaches by essentially making a chain of calls where the output is chained and linked. This allowed us to not only find the missing tasks but then make an action plan personalized.

Challenges we ran into

The biggest challenge on this project was time management and frontend. All of our team-members work full-time and as such, finding time to work on this project between our work and personal lives was a very difficult task. In addition to time management difficulties, configuring the API response from the backend to look easy to use and intuitive on the front-end was much more harder of a task than we all anticipated.

Accomplishments that we're proud of

We are very proud of churning out an end-to-end application that displayed the capabilities of our MVP in a very short time frame. Accomplishing this was reduced times and schedules make us even more proud. In addition, we believe the UI/UX of our application lends itself very well to usage by international students.

What we learned

We learned a lot about time management and how to work with different schedules. We also learnt tons about the criticality of configuring with frontend to a stable backend. In addition, we experimented with various techniques with the Gemini and that widened our scope that we can accomplish with this powerful technology. For example, some of our data that we use came from using the Gemini's VLM in the Google AI Studio - allowing us to scrape data really fast.

What's next for Borderless

There are no borders for Borderless. Simply put, the scope for expanding this to other visa categories already presents a massive value addition. Furthermore, we hope to expand on the capabilities of the application with improving the checklist generation by having a robust RAG pipeline, implementing VLMs to enhance understanding of the legal forms, and having a system-instructed chatbot that allows the student to ask any and all questions.

Built With

Share this project:

Updates