WiseWorks

Dashboard
Edit Profile
Recommendation Page

Inspiration Canada is facing a 300,000+ skilled trades shortage. The workers exist. The demand exists. The infrastructure connecting them is completely broken. Every AI hiring tool ever built was designed for HR departments & for people with LinkedIn profiles, polished resumes, and keyword-optimized PDFs. Nobody built anything for Carlos. Carlos is a Red Seal electrician with 20 years of experience. He's never written a resume. He finds work through word of mouth. He has no idea he's being underpaid by $10 an hour. We spotted a pattern nobody else in that building was looking at: the talent exists, the demand exists, and the system was never built for the worker. WiseWorks is our answer.

What It Does WiseWorks is a voice-first AI career agent built exclusively for skilled trades workers in Canada. No resume. No LinkedIn. No forms. Just speak.

🎙 Voice Onboarding - Worker speaks their experience naturally. ElevenLabs converts speech to text. Moorcheh AI structures it into a verified skills profile in under 10 seconds. 💰 Pay Fairness Detection - Every job posting is analyzed against real Canadian market wage data. If a job is underpaying you, WiseWorks tells you exactly how much and what to negotiate for. 🎯 AI Job Matching - Moorcheh semantic search matches your actual skills, certifications, and location against live job postings with ranked match scores. 📋 Contract Explainer - Paste any employment contract clause. Get a plain-language explanation with key points, risks, and recommendations. No lawyers needed. 💬 AI Chat Assistant - Ask anything about jobs, wages, or contracts. Powered by Moorcheh RAG over real worker profiles and job data. 📄 Resume Generator - Auto-generates a professional, downloadable PDF resume from your voice profile.

How We Built It Backend: FastAPI (Python) with SQLite for persistent worker profiles. Deployed locally with uvicorn. AI / RAG: Moorcheh AI powers everything intelligent - semantic search across worker profiles, job postings, and Canadian wage data namespaces. Direct AI mode handles pay analysis, contract explanation, and profile extraction from voice transcripts. Voice: ElevenLabs STT as the primary transcription engine with Google Gemini as a fallback. The onboarding asks structured questions one at a time - worker speaks each answer, ElevenLabs transcribes it, Moorcheh structures the full profile. Frontend: React + Vite with a custom dark teal/gold design system inspired by our WiseWorks owl logo. Particle network background, custom cursor, Cinzel typography. Built to look nothing like a generic AI hackathon project. Data: Pre-loaded Canadian trades job postings and Labour Market wage data (2025-2026) across Moorcheh namespaces for real-time pay fairness analysis.

Challenges We Ran Into The hardest technical challenge was getting Moorcheh's job matching to respond to a specific worker rather than all workers in the knowledge base. The naive approach returned results for every worker in the namespace. The fix was fetching the worker's profile directly from SQLite first, then passing that text as context to Moorcheh's Direct AI mode - bypassing the namespace query entirely for personalized matching. ElevenLabs transcribes numbers as words - "thirty eight" instead of "38". We built a word-to-number parser to handle this across all numeric fields in onboarding. Python 3.13 compatibility with pydantic v1 was a surprise blocker that cost us an hour. We resolved it by migrating to pydantic v2 with field_validator syntax. The biggest non-technical challenge: constantly asking ourselves - would someone who has never used a smartphone be able to use this? That question shaped every design decision.

Accomplishments We're Proud Of

End-to-end voice-to-profile pipeline working in under 10 seconds Moorcheh pay fairness analysis returning real market data with actionable recommendations A UI that genuinely looks like a product - not a hackathon project Building for a demographic that every other team in that building ignored Shipping a fully working product in 36 hours with a team of 3

What We Learned We learned that the hardest part of a hackathon isn't the technology. It's finding a problem worth solving. We spent the first hour debating ideas. Nothing felt meaningful. Then we looked at the data - 300,000+ unfilled trades positions, a workforce aging out of the digital job market, and zero AI tools built with them in mind. That's when WiseWorks became real. We also learned that Moorcheh's RAG capabilities are genuinely powerful when used correctly - the semantic search across wage data for pay fairness analysis was one of the most satisfying moments of the build.

What's Next for WiseWorks

Multilingual support - Punjabi, Tagalog, Portuguese, Spanish. The actual languages of Canada's trades workforce. Mobile-first PWA - Optimized for workers on job sites, not at desks. Red Seal certification verification - Direct API integration with provincial certification bodies. Employer portal - Let employers search verified worker profiles directly. Expand beyond trades - Every blue collar worker in Canada deserves this tool.