OpenSolve

A platform aggregating real funded problems (YC Requests for Startups, Indian government innovation challenges, industry-posted problems). Students submit public, verifiable solutions. Organizations view ranked submissions, see student profiles, and contact top performers.

Built as an Open Innovation Platform prototype.

Tech Stack

• Next.js 14 (App Router)
• React, Tailwind CSS, Framer Motion
• AWS DynamoDB (AWS SDK v3)
• TypeScript

Setup & Running Locally

1. Install dependencies:

   npm install

2. Set up environment variables (.env.local):

   # AWS Credentials for DynamoDB
   AWS_REGION=us-east-1
   AWS_ACCESS_KEY_ID=your_key
   AWS_SECRET_ACCESS_KEY=your_secret
   
   # Table Names (Optional, defaults exist)
   DYNAMODB_TABLE_PROBLEMS=OpenSolve_Problems
   DYNAMODB_TABLE_SUBMISSIONS=OpenSolve_Submissions
   DYNAMODB_TABLE_ORGANIZATIONS=OpenSolve_Organizations
   DYNAMODB_TABLE_QATHREADS=OpenSolve_QAThreads
   DYNAMODB_TABLE_PROFILES=OpenSolve_Profiles

3. Run the application:

   npm run dev

4. Seed the database with mock data (if running a local DynamoDB instance or wanting to populate a dev table):

   node scripts/seed.js

Architecture & Infrastructure

• Auth: Production-ready authentication using Clerk, including role-based access control (RBAC) via middleware.
• Database: Strictly adheres to DynamoDB Single-Table Design patterns. All primary access paths utilize QueryCommand via Global Secondary Indexes (GSIs) to ensure $O(1)$ scaling.
• Rate Limiting: Uses Upstash Redis (@upstash/ratelimit) for distributed edge-compatible rate limiting on user-facing routes, failing open if misconfigured to ensure availability.
• Security: Robust middleware protecting API routes and enforcing deadline constraints. Score evaluation logic is securely isolated in backend API flows.

Automated Ingestion

OpenSolve includes an automated problem-ingestion scraper pipeline designed to run as a background Vercel Cron Job.

• Schedule: Runs daily at midnight (0 0 * * *). This complies with Vercel's Hobby tier limits and avoids aggressively hitting source sites.
• Scrapers:
  • YC Requests for Startups
  • UK Innovation Funding
  • SPRIND
• Deduplication: The scraper uses a sourceUrl-index Global Secondary Index in DynamoDB to check if a problem has already been ingested, preventing duplicates.
• Verification: Ingested problems are explicitly marked as verified: false and attributed to an aggregated system user. Organizations can claim their listings to add the verified badge.
• Security: The ingestion endpoint (/api/cron/ingest) is protected via the CRON_SECRET environment variable, which Vercel Cron Jobs securely pass as a Bearer token.
• Limitations: HTML scraping is inherently fragile. If the source sites update their layout significantly, the cheerio parsers will gracefully log an error and skip ingestion rather than crashing. A more robust production version would utilize official APIs when available.