OppTracker

Inspiration

Every year, millions of students and young professionals apply to international fellowships, internships, hackathons, and funded programs. The process is scattered across dozens of websites, deadlines get missed, generic cover letters take hours to draft from scratch, and fraudulent programs waste valuable time and money. We built OppTracker to turn this chaotic journey into a structured, automated, and secure workflow.

What it does

OppTracker is an all-in-one opportunity management platform powered by AI:

• Smart Dashboard & Pipeline: Visualizes your entire pipeline across 7 distinct workflow stages (Need to Apply, Applied, Under Review, Interview, Accepted, Rejected, and Scam) with auto-countdown deadline tracking and browser notifications.
• AI URL Analyzer: Extracts deadlines, funding types, requirements, and benefits from any program link in seconds.
• AI Cover Letter Generator: Generates 300–400 word customized cover letters based on the opportunity details and the user's stored profile.
• Scam Detector & Blacklist: Evaluates fraud signals with a 0–100 risk score, flags red/green flags, and provides clear recommendations.
• Smart Chat Assistant: A conversational assistant that understands your full database of opportunities to answer queries and prioritize upcoming deadlines.

How I built it

• Frontend: Built with React 19, TypeScript, and Vite for a fast, responsive user interface styled with Tailwind CSS 4 in a custom dark-mode theme.
• Backend & Database: Powered by Supabase (PostgreSQL) utilizing Row Level Security (RLS) and database CHECK constraints to enforce status workflow integrity.
• AI Layer: Integrated Google Gemini 3.6 Flash to handle instant text extraction, cover letter generation, contextual chat assistance, and scam heuristic evaluation.
• Deployment: Hosted with zero-latency global edge delivery on Vercel.

Challenges I ran into

• Strict State Transitions: Designing a database schema and UI state machine that cleanly enforces the 7-stage workflow without allowing invalid state jumps.
• Reliable Information Extraction: Fine-tuning prompts for Gemini to consistently extract unstructured web page data (deadlines, funding status, criteria) into predictable JSON structures.
• Scam Scoring Heuristics: Balancing fraud-detection criteria so real opportunities aren't falsely flagged while catching subtle red flags like suspicious fee requests or non-verifiable organizers.

Accomplishments that I'm proud of

• Building a full-stack, AI-native platform with zero ongoing hosting costs by utilizing modern free tiers.
• Creating an interface that handles full CRUD management, intelligent analysis, and proactive browser reminders in one seamless dashboard.
• Developing a practical tool ready for immediate real-world use by students and job seekers worldwide.

What I learned

• Integrating Google Gemini directly into modern React workflows for fast client-side intelligence.
• Enforcing bulletproof data integrity using Supabase RLS and custom PostgreSQL constraints.
• Structuring efficient prompt pipelines for real-time document analysis and contextual chat assistants.

What's next for OppTracker

• Browser Extension: One-click opportunity capture directly from LinkedIn, Twitter, and program websites.
• Calendar Sync: Two-way synchronization with Google Calendar and Outlook.
• Multi-User Collaboration: Shared pipelines and community-driven scam reporting feeds.
• Mobile App: A React Native companion app for tracking on the go.

Built With

• css3
• google-gemini
• html5
• javascript
• postgresql
• react
• supabase
• tailwind-css
• typescript
• vercel
• vite