This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle.
What I Built
Finding a research internship is harder than it should be.
When I started exploring research opportunities, I realized how much time students spend searching faculty websites, finding relevant professors, reading papers, writing cold emails, and tracking applications.
Most students end up with dozens of browser tabs, messy spreadsheets, and generic email templates.
So I built WRID (Winter Research Internship Desk) — a workspace that brings the entire research internship process into one place.
Here's what it does:
Faculty Directory: Search 1,610+ professors across 49 institutions, with filters for departments and research interests.
AI Resume Matching: Upload your resume and find professors whose research aligns with your skills and projects.
Academic Email Studio: Use Gemma and Llama to draft, personalize, and review research outreach emails.
Fellowship Discovery: Explore research programs at institutions such as IITs, MIT, ETH Zürich, and CERN.
Application Tracker: Manage applications, follow-ups, and progress without maintaining separate spreadsheets.
Research Roadmap: A structured six-week plan to prepare for research internships.
The idea is simple: spend less time figuring out whom to contact and more time doing actual research.
Who Would Pay for It?
Students applying for competitive research internships are the primary users. A potential paid tier could offer advanced faculty matching, deeper research analysis, and additional AI-assisted outreach features.
The core discovery tools should remain accessible.
Demo
Live Website:(https://hacktoberfest-winter-research.vercel.app)
Code
GitHub:(https://github.com/CoderOggy78/WRID)
How I Built It
WRID uses React 19, TypeScript, Vite, and Tailwind CSS for the frontend.
For AI, I integrated Google Gemma and Meta Llama 3.3 70B through Groq into the academic writing workflow.
The system combines two approaches:
Rule-based matching: Extracts skills, projects, and research interests from resumes and compares them against faculty research profiles. The resulting scores indicate estimated alignment, not acceptance chances.
Open-weight LLMs: Help students refine cold emails, improve academic tone, and prepare personalized follow-ups.
Resume text extraction and application tracking run client-side, with application data stored in the browser. AI requests using hosted providers are processed through their respective APIs.
The biggest challenge was making the recommendations useful rather than simply displaying a list of professors with arbitrary scores.
Why Does Open Innovation Matter?
I didn't want WRID's AI features to depend entirely on one proprietary model.
Using open-weight models like Gemma and Llama gives me the flexibility to experiment with different models, compare their performance, and potentially support local inference in the future.
That matters for a student-focused product where cost, accessibility, and privacy are important.
It also means the AI layer can evolve without rebuilding the entire application.
The goal isn't to add AI everywhere. It's to use it where it genuinely reduces the work students have to do.
Prize Categories
Side Hustle — Open-Source AI Challenge, Week 2
WRID is still evolving, and there's plenty I want to improve, particularly research matching, faculty data quality, and fellowship discovery.
If you're a student applying for research internships, I'd genuinely appreciate you trying it.
What part of the research internship application process do you find most frustrating?
I'd love to use that feedback to decide what to build next.
Top comments (0)