Back to Jaanvi's profile
10views

HireCampus

Turning campus placements into a smarter journey with AI-powered career preparation, automated placement management, and student readiness tracking.

Jaanvi Nayak
  • Problem, approach and result explained
  • Public source code
  • Team confirmed roles· 4 of 4
  • Expert reviewed· 5 reviews
HireCampus

—

Integrated layers: Frontend, Backend & Database

—

Laravel database migrations

—

Locally integrated LLM – Llama 3.2 via Ollama

Overview

The problem

The problem is for the students who are in colleges, the placement officers, and the recruiters. the problem matters a lot for all of them as a student always does not know what and how to prepare for jobs, and for the placement officers they have too much burden on the time of placement this simplifies their work, recruiters will have it easy to identify their required student. Our website helps the students to analyze their skill-gaps and what they are lacking for the job they want to apply for, they can practice for the interviews and quiz. For placement officers they can easily shortlist the students and can see their progress. For recruiters our website does not let them carry all the burdens and they can easily see which student is more eligible for their work.

The approach

CampusLink was developed to address the gap between university placement processes and students' actual career readiness. Instead of building another basic placement portal, we designed an integrated platform connecting students, Training and Placement Officers (TPOs), recruiters, and college administrators.

We used Next.js, React, TypeScript, and Tailwind CSS for the frontend, Laravel for backend APIs and business logic, and PostgreSQL for database management. We integrated Ollama with Llama 3.2 to introduce locally hosted AI capabilities.

Our approach focused on building practical modules, including AI-powered resume assistance, interview preparation, career-readiness assessment, placement journey tracking, and conflict-aware placement scheduling. We also designed an assessment workflow in which administrators can initiate assessments, TPOs verify them, and students receive the approved assessments.

We prioritized modular architecture, role-based access control, secure authentication, and reliable database integration to ensure that the platform could support real institutional workflows.

The result

During the hackathon, we implemented a Next.js frontend, Laravel backend, and PostgreSQL database, with the application deployed and connected to its backend infrastructure. We integrated Ollama with Llama 3.2 and verified local AI connectivity and text generation through our existing Laravel AI service.

The platform includes role-based authentication, AI-assisted resume building, interview preparation, career-readiness scoring, placement journey tracking, assessment management, and conflict-aware placement scheduling.

We validated key technical components through API response checks, AI generation tests, database integration checks, and backend testing. For example, our authentication endpoint returned a successful HTTP 200 response, and our Ollama integration successfully generated responses through Laravel's AI manager.

Rather than claiming unsupported performance statistics, we focused on demonstrating working integrations and functional workflows. CampusLink transforms fragmented placement activities into a centralized, structured, and AI-assisted ecosystem, helping institutions organize recruitment while enabling students to prepare more effectively for their careers.

AI shortlisting

Reflection

One of our biggest challenges was integrating multiple components into a single working system within the limited hackathon timeframe. Connecting the frontend, Laravel backend, PostgreSQL database, and locally hosted AI model required careful coordination and debugging.

We also learned that implementing AI is not simply about connecting a language model to an application. It requires meaningful use cases, proper data handling, reliable backend integration, and responsible access to student information.

Working on CampusLink helped us understand real-world software development practices, including API integration, database management, authentication, deployment, and teamwork under time constraints.

If we had more time, we would improve the AI capabilities, expand recruiter-side functionality, strengthen automated testing, and introduce more advanced analytics to help institutions understand student career readiness.

Most importantly, this project taught us to approach technology from the perspective of actual users and solve problems through practical implementation rather than just ideas.

Submitted to CAMPUSLINK: Placement Season Without the Spreadsheet Tandav

Your idea in one line

An AI-driven campus placement platform that seamlessly connects students, recruiters, and placement officers to orchestrate the entire hiring lifecycle

What will be working by the finale?

1. A conflict-aware drive scheduling mechanism plus automated notification workflow. 2. AI resume builder. 3. Secure AI examination and interview mode. 4. Student placement journey tracker. 5. AI Powered communication skill analysis.

Tech stack and hardware you plan to use

1. The Software Layer (Next.js, Laravel, MySQL) 2. The AI Layer (Ollama, LLaMA 3.2) 3. The Hardware/Infrastructure (AWS t4g.small EC2 instance)

What have you built or tested so far?

We have built the core platform featuring role-based dashboards for students, recruiters, and placement officers. We integrated LLaMA 3.2 for automated resume parsing, built a dynamic readiness scoring engine, and deployed an interactive AI navigation assistant. So far, we have successfully tested the AI validated the readiness scoring logic, and verified the end-to-end placement drive workflows.

Pitch deck (PDF)
Pitch deck (PDF)PDF
What's working now?

Working: The end-to-end placement pipeline—including authentication, student profiles, the 4-round selection workflow with video interviews, offer management, and institutional analytics—is fully operational and verified by 342+ passing tests. Partly Working: Code sandboxing, dynamic AI completions, and internal chat operate using simulated mock adapters, local Ollama/rule fallbacks, and REST polling rather than isolated Docker runners or WebSockets. Not Reached: The dedicated Mentor portal, historical statistical predictive ML models

Deck or report (PDF)
Deck or report (PDF)PDF
A photo or screenshot
A photo or screenshot: HireCampus

AI tools used

ChatGPTGemini

Help in building the website

Links & files

Artifacts

1

Gallery

5