AI Interviewer by LiveKit
Activities Page
List of colleges, requirements, deadlines
Multi agent system for essays
Betting Market (Under construction)
College requirements, stats, essays (direct link to essays page)

About the Project

This year felt tedious and exhausting.

Between writing college supplemental essays, fixing activity descriptions, and optimizing honors titles, the college application process became overwhelming. College advisors exist to help with this, but over time this support has turned into a growing billion-dollar industry, with some charging upwards a thousand an hour. Prices have risen to the point where many students have a hard time paying for expert guidance at all.

That gap is what inspired AdmitX.

As students who went through the process ourselves, we realized the problem was not motivation or effort, but access. We wanted to give students real guidance without the financial barriers that dominate college admissions today. AdmitX is a company for students by students.

AdmitX is an AI-powered college advising platform that provides personalized guidance, structured feedback, and real-time support throughout the admissions process. Instead of paying thousands of dollars for private advisors, students can access tailored advice that adapts to their goals, background, and progress for more affordable rates.

How We Built It

AdmitX is a full-stack application built using React, TypeScript, and Python, supported by a backend that connects multiple AI and infrastructure services through custom API endpoints. These endpoints allow all services to run inside a single server environment, enabling fast iteration and reliable communication between systems.

We used DevSwarm to coordinate multiple AI agents during development. This allowed us to debug efficiently, reason through architectural decisions, and solve complex issues under tight time constraints. Instead of debugging in isolation, we could parallelize reasoning and rapidly converge on solutions.

To support real-time advising, we integrated LiveKit to build a low-latency voice agent. This made interactions feel natural and conversational rather than scripted. We paired this with Groq for fast inference and Supabase for authentication and data storage.

At the core of AdmitX is an AGI-first product. Rather than building static forms or rule-based flows, we designed the system to reason about student inputs dynamically and adapt guidance in real time suited for each personal application.

Technical Challenges

Most of our challenges appeared when everything had to work together.

We encountered failures caused by mismatched API request and response formats, authentication state issues that broke downstream calls, and timing problems during LiveKit session initialization. At one point, inference requests returned correctly from Groq but failed silently when passed through the server, making debugging especially difficult.

Solving these problems required restructuring endpoints, improving logging, and simplifying how data flowed through the system. We continuously tested integrations to ensure reliability during live demos. DevSwarm played a critical role in isolating failures and helping us move quickly without sacrificing stability.

What We Learned

Through building AdmitX, we learned how to:

● Design scalable API architectures across multiple services

● Build real-time voice applications with low latency using LiveKit

● Debug distributed systems under pressure when multiple tools must work together

● Integrate fast AI inference into production-style workflows using tools like Groq

● Use platforms like Supabase for authentication and data handling in a live application

● Leverage DevSwarm to parallelize reasoning, debugging, and architectural decision-making

● Prioritize reliability and stability during live demos

More importantly, we learned how to turn a personal problem into a scalable product.

Impact and Scalability

AdmitX addresses a real and growing problem. Millions of students apply to college every year, and personalized advising is becoming increasingly inaccessible. By lowering costs and increasing access, AdmitX has the potential to reach students at a global scale.

Because AdmitX is software-first, it scales far beyond individual users. The platform can expand into schools, counselors, and district-level education systems through licensing, subscriptions, and institutional partnerships, creating strong, recurring revenue. At the same time, this growth directly benefits communities by increasing access to high-quality admissions guidance, especially for students who lack personalized support. Built to scale well beyond a hackathon, AdmitX is positioned as a sustainable startup with long-term financial viability and meaningful social impact.

Why It Matters

The college admissions process should reward effort and potential for the hard work that students have put in throughout high school, not access to money.

AdmitX exists to close that gap. This project represents what is possible when students build practical, AI-driven solutions for problems they have lived through themselves.