Documentation
Title: HULUM - AI Opportunity Decision Agent
The Problem: Ethiopian students and young professionals face opportunity overload with zero eligibility intelligence. Thousands of scholarships, fellowships, internships, and jobs are scattered across Telegram channels, Facebook groups, and websites every week but there is no structured way to answer the only question that matters: "Am I actually eligible for this, and is it worth my time to apply?" Without guidance, students miss opportunities they could have won, waste hours on applications they never qualified for, and fall into decision paralysis.
The Solution:
HULUM introduces a "Decision-First" workflow. It saves a structured user profile (education, skills, experience, languages, interests) and compares it against any opportunity using a deterministic TypeScript engine not an LLM guess — to return one clear verdict: APPLY, MAYBE, or SKIP. It then explains why, surfaces every unmet requirement, and generates a concrete readiness checklist so users know exactly what to improve before the next cycle.
Key Highlights:
• Deterministic Decision Engine: APPLY / MAYBE / SKIP is computed by TypeScript rules zero hallucination risk on the verdict that matters most.
• Gap Analysis & Readiness Checklist: Every requirement is marked MET, NOT_MET, or UNCERTAIN with a concrete to-do list for the next application cycle.
• Screenshot Analyze: Users paste or upload any opportunity screenshot Groq Vision extracts the facts, then decide() runs against the profile instantly.
• Ask HULUM (AI Assistant): Text and voice interface powered by Groq, with a Voxide microphone button and graceful fallback to typed input.
• Verified Opportunity Catalog: Sourced by a background investigator worker from evidenced pages — never invented. Every listing links to its official source.
Submission Link: https://hulum-alpha.vercel.app
Repo Link: https://github.com/mumekonin/hulum
