StudyNest

The idea began in my first weeks at university. I knew nobody. I wanted someone taking the same subjects, in the same year, who studied the way I did. Apps for messaging people were everywhere. None of them helped me find the right people.

After the remote learning period, many students ended up studying alone. Nothing brought together the three things they needed: finding a suitable study partner, a space to work with them, and help the moment a problem came up while studying. Those are the moments a student needs support.

I started in June 2025, during a summer course, a week before midterms. The capstone was not due to begin for months. I started early on purpose, to build the project calmly instead of under the pressure of a full semester.

I wrote down three ideas, picked one, and had a working prototype three days later. I kept building on the same prototype for a full year, up to the final capstone evaluation. I presented the work there for the Information Technology Department, and entered the project into the London International Youth Science Forum.

Moodle and Google Classroom were never the competition. They manage courses and organise teaching material, and they do the job well.

Neither one helps a student find people to study with, or work alongside them once found. StudyNest completes those platforms and adds the social, collaborative side they leave out.

I did not want to leave a student searching through hundreds of profiles.

The engine scores compatibility across the profile: subjects, academic level, study style, free time, university, country, age, interests and past experience. The list comes back ordered by best fit. The weights are not equal. Subjects carry 35% of the match score, academic level 20%, and study style and free time 15% each.

A search tells you who meets the conditions. The ranking tells you who is worth starting with.

Finding a study partner is not enough on its own, so the platform gives students a space to study in directly.

Rooms carry live chat, voice and video calls, screen sharing, and scheduled study sessions. They support host, co-host, participant and viewer roles, which keeps a session orderly as the number of people grows.

Most of the work went into the details nobody notices. When the host leaves, control passes automatically to the co-host, then to the first participant who joined. Clients send a connection signal every 15 seconds, and anyone who drops gets 30 seconds to return to the same session without losing their state or starting again.

StudyNest includes a Gemini-powered assistant. The assistant answers typed questions, reads the documents and images a student uploads, explains the contents, and replies by text or voice.

The assistant also turns a course syllabus into an organised study plan, or into flashcards for revision.

The information was always available. What was missing was reaching the right information at the right time.

The system handles personal and academic data, so security was a core part of the project rather than an afterthought.

The work covers JWT authentication with refresh-token rotation, role-based permissions, two-factor authentication and email confirmation, alongside 125 passing tests across 20 data models.

An access token lives 15 minutes and a refresh token seven days, which limits the damage a stolen token does. The platform disables accounts logically instead of deleting them, so a student returns later on the same email without losing their data.

Forty-four students used the platform during the pilot.

The surveys showed 86% were satisfied with the experience, while 95% said the AI assistant improved how they study.

The project was graded Distinction as a senior capstone, and selected among the top 30 of 738 projects nationally at the London International Youth Science Forum finals held in Muscat.