UpScore — AI Upwork Job Matcher

UpScore is a SaaS platform that ingests Upwork job listings at scale, runs asynchronous AI analysis and rule-based scoring, and surfaces the highest-fit opportunities to freelancers — reducing manual filtering from hours to minutes.

Freelancers spend significant time manually scanning Upwork for jobs that match their skills and rate. The volume of listings makes it easy to miss high-fit opportunities.

Full-stack development — architecture, data ingestion, AI analysis pipeline, authentication, and frontend.

Built an automated ingestion pipeline on a VPS that pulls job data via an Upwork MCP client connected to an Upwork MCP server, stores it in Postgres via Supabase, and runs asynchronous AI analysis with rule-based scoring to rank opportunities by fit.

Design details for this case study are being prepared. The interface was built with a focus on clarity, performance, and a clean user experience.

The platform was built with the following technical components:

• 01VPS data ingestion
• 02Upwork MCP client / server
• 03Vercel
• 04Supabase
• 05Postgres
• 06Job API
• 07Asynchronous AI analysis
• 08Rule-based scoring
• 09Upwork OAuth / PKCE
• 10Encrypted token storage
• 11OAuth proxy
• 12Rate limiting
• 13CSP
• 14Supabase migrations

Detailed challenges and how they were solved will be added here. The project involved handling rate limiting, secure token storage, and asynchronous processing at scale.

Outcome metrics for this project will be documented here. The platform is live and in use.